Merge branch '3.0' into feat/sangshuduo/TD-14141-update-taostools-for3.0

docs/en/12-taos-sql/10-function.md

@@ -273,7 +273,7 @@ SELECT CONCAT(str1|column1, str2|column2, ...) FROM { tb_name | stb_name } [WHER
 
 **Return value type**: If all input strings are VARCHAR type, the result is VARCHAR type too. If any one of input strings is NCHAR type, then the result is NCHAR. If input strings contain NULL value, the result is NULL. 
 
-**Applicable data types**: VARCHAR, NCHAR.  At least 2 input strings are requird, and at most 8 input strings are allowed. 
+**Applicable data types**: VARCHAR, NCHAR.  At least 2 input strings are required, and at most 8 input strings are allowed. 
 
 **Applicable table types**: table, STable.
 
@@ -290,7 +290,7 @@ SELECT CONCAT_WS(separator, str1|column1, str2|column2, ...) FROM { tb_name | st
 
 **Return value type**: If all input strings are VARCHAR type, the result is VARCHAR type too. If any one of input strings is NCHAR type, then the result is NCHAR. If input strings contain NULL value, the result is NULL. 
 
-**Applicable data types**: VARCHAR, NCHAR. At least 3 input strings are requird, and at most 9 input strings are allowed. 
+**Applicable data types**: VARCHAR, NCHAR. At least 3 input strings are required, and at most 9 input strings are allowed. 
 
 **Applicable table types**: table, STable.
 
@@ -527,6 +527,7 @@ SELECT TIMEDIFF(ts1 | datetime_string1, ts2 | datetime_string2 [, time_unit]) FR
 - Time unit specified by `time_unit` can be:
   1b(nanosecond), 1u(microsecond),1a(millisecond),1s(second),1m(minute),1h(hour),1d(day),1w(week).
 - If `time_unit` parameter is not specified, the precision of the returned time duration is same as the precision set for the current database in use.
+- If input date-time string cannot be converted to UNIX timestamp, NULL value is returned.
 
 #### TIMETRUNCATE
 
@@ -547,6 +548,7 @@ SELECT TIMETRUNCATE(ts_val | datetime_string | ts_col, time_unit) FROM { tb_name
 - Time unit specified by `time_unit` can be:
   1b(nanosecond),1u(microsecond),1a(millisecond),1s(second),1m(minute),1h(hour),1d(day),1w(week).
 - The precision of the returned timestamp is same as the precision set for the current database in use.
+- If input date-time string cannot be converted to UNIX timestamp, NULL value is returned.
 
 #### TIMEZONE
 
@@ -599,7 +601,7 @@ FROM { tb_name | stb_name } [WHERE clause]
 
 **Return value type**: DOUBLE.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -654,7 +656,7 @@ SELECT ELAPSED(field_name[, time_unit]) FROM { tb_name | stb_name } [WHERE claus
 
 **Return value type**：DOUBLE.
 
-**Applicable Column type**：TIMESTAMP.
+**Applicable data type**：TIMESTAMP.
 
 **Applicable tables**: table, STable, outter in nested query.
 
@@ -693,7 +695,7 @@ SELECT SPREAD(field_name) FROM { tb_name | stb_name } [WHERE clause];
 
 **Return value type**: DOUBLE.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -709,7 +711,7 @@ SELECT STDDEV(field_name) FROM tb_name [WHERE clause];
 
 **Return value type**: DOUBLE.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -723,7 +725,7 @@ SELECT SUM(field_name) FROM tb_name [WHERE clause];
 
 **Return value type**: DOUBLE.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -737,7 +739,7 @@ SELECT HYPERLOGLOG(field_name) FROM { tb_name | stb_name } [WHERE clause];
 
 **Return value type**: INTEGER.
 
-**Applicable column types**: All data types.
+**Applicable data types**: All data types.
 
 **More explanations**: The benefit of using hyperloglog algorithm is that the memory usage is under control when the data volume is huge. However, when the data volume is very small, the result may be not accurate, it's recommented to use `select count(data) from (select unique(col) as data from table)` in this case.
 
@@ -751,14 +753,14 @@ SELECT HISTOGRAM(field_name，bin_type, bin_description, normalized) FROM tb_nam
 
 **Return value type**：DOUBLE or BIGINT, depends on normalized parameter settings.
 
-**Applicable column type**：Numerical types.
+**Applicable data type**：Numerical types.
 
 **Applicable table types**: table, STable.
 
 **Explanations**：
 
-1. bin_type: parameter to indicate the bucket type, valid inputs are: "user_input", "linear_bin", "log_bin"。
-2. bin_description: parameter to describe the rule to generate buckets，can be in the following JSON formats for each bin_type respectively:
+- bin_type: parameter to indicate the bucket type, valid inputs are: "user_input", "linear_bin", "log_bin"。
+- bin_description: parameter to describe the rule to generate buckets，can be in the following JSON formats for each bin_type respectively:
 
    - "user_input": "[1, 3, 5, 7]": User specified bin values.
 
@@ -776,7 +778,7 @@ SELECT HISTOGRAM(field_name，bin_type, bin_description, normalized) FROM tb_nam
      "infinity" - whether to add（-inf, inf）as start/end point in generated range of bins.
      The above "log_bin" descriptor generates a set of bins:[-inf, 1.0, 2.0, 4.0, 8.0, 16.0, +inf].
 
-3. normalized: setting to 1/0 to turn on/off result normalization.
+- normalized: setting to 1/0 to turn on/off result normalization.
 
 ### PERCENTILE
 
@@ -788,7 +790,7 @@ SELECT PERCENTILE(field_name, P) FROM { tb_name } [WHERE clause];
 
 **Return value type**: DOUBLE.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table.
 
@@ -808,7 +810,7 @@ SELECT BOTTOM(field_name, K) FROM { tb_name | stb_name } [WHERE clause];
 
 **Return value type**: Same as the column being operated upon.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -828,7 +830,7 @@ SELECT FIRST(field_name) FROM { tb_name | stb_name } [WHERE clause];
 
 **Return value type**: Same as the column being operated upon.
 
-**Applicable column types**: All data types.
+**Applicable data types**: All data types.
 
 **Applicable table types**: table, STable.
 
@@ -848,7 +850,7 @@ SELECT INTERP(field_name) FROM { tb_name | stb_name } [WHERE where_condition] [
 
 **Return value type**: Same as the column being operated upon.
 
-**Applicable column types**: Numeric data types.
+**Applicable data types**: Numeric data types.
 
 **Applicable table types**: table, STable, nested query.
 
@@ -872,7 +874,7 @@ SELECT LAST(field_name) FROM { tb_name | stb_name } [WHERE clause];
 
 **Return value type**: Same as the column being operated upon.
 
-**Applicable column types**: All data types.
+**Applicable data types**: All data types.
 
 **Applicable table types**: table, STable.
 
@@ -892,7 +894,7 @@ SELECT LAST_ROW(field_name) FROM { tb_name | stb_name };
 
 **Return value type**: Same as the column being operated upon.
 
-**Applicable column types**: All data type.
+**Applicable data types**: All data type.
 
 **Applicable table types**: table, STable.
 
@@ -911,7 +913,7 @@ SELECT MAX(field_name) FROM { tb_name | stb_name } [WHERE clause];
 
 **Return value type**: Same as the data type of the column being operated upon.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -925,7 +927,7 @@ SELECT MIN(field_name) FROM {tb_name | stb_name} [WHERE clause];
 
 **Return value type**: Same as the data type of the column being operated upon.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -939,7 +941,7 @@ SELECT MODE(field_name) FROM tb_name [WHERE clause];
 
 **Return value type**:Same as the data type of the column being operated upon.
 
-**Applicable column types**: All data types.
+**Applicable data types**: All data types.
 
 **More explanations**:Considering the number of returned result set is unpredictable, it's suggested to limit the number of unique values to 100,000, otherwise error will be returned.
 
@@ -976,7 +978,7 @@ SELECT TAIL(field_name, k, offset_val) FROM {tb_name | stb_name} [WHERE clause];
 
 **Return value type**: Same as the column being operated upon.
 
-**Applicable column types**: All data types.
+**Applicable data types**: All data types.
 
 ### TOP
 
@@ -988,7 +990,7 @@ SELECT TOP(field_name, K) FROM { tb_name | stb_name } [WHERE clause];
 
 **Return value type**: Same as the column being operated upon.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -1008,7 +1010,7 @@ SELECT UNIQUE(field_name) FROM {tb_name | stb_name} [WHERE clause];
 
 **Return value type**: Same as the column or tag being operated upon.
 
-**Applicable column types**: All data types.
+**Applicable data types**: All data types.
 
 **More explanations**:
 
@@ -1050,7 +1052,7 @@ SELECT DERIVATIVE(field_name, time_interval, ignore_negative) FROM tb_name [WHER
 
 **Return value type**: DOUBLE.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -1070,7 +1072,7 @@ SELECT {DIFF(field_name, ignore_negative) | DIFF(field_name)} FROM tb_name [WHER
 
 **Return value type**: Same as the column being operated upon.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -1090,7 +1092,7 @@ SELECT IRATE(field_name) FROM tb_name WHERE clause;
 
 **Return value type**: DOUBLE.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 
@@ -1183,7 +1185,7 @@ SELECT TWA(field_name) FROM tb_name WHERE clause;
 
 **Return value type**: DOUBLE.
 
-**Applicable column types**: Numeric types.
+**Applicable data types**: Numeric types.
 
 **Applicable table types**: table, STable.
 

docs/zh/12-taos-sql/10-function.md

@@ -401,7 +401,7 @@ SELECT CAST(expression AS type_name) FROM { tb_name | stb_name } [WHERE clause]
 
 **返回结果类型**：CAST 中指定的类型（type_name)。
 
-**适用数据类型**：输入参数 expression 的类型可以是BLOB、MEDIUMBLOB和JSON外的所有类型。
+**适用数据类型**：输入参数 expression 的类型可以是除JSON外的所有类型。
 
 **嵌套子查询支持**：适用于内层查询和外层查询。
 
@@ -410,10 +410,10 @@ SELECT CAST(expression AS type_name) FROM { tb_name | stb_name } [WHERE clause]
 **使用说明**：
 
 - 对于不能支持的类型转换会直接报错。
-- 对于类型支持但某些值无法正确转换的情况对应的转换后的值以转换函数输出为准。目前可能遇到的几种情况：
+- 对于类型支持但某些值无法正确转换的情况，对应的转换后的值以转换函数输出为准。目前可能遇到的几种情况：
         1）字符串类型转换数值类型时可能出现的无效字符情况，例如"a"可能转为0，但不会报错。
         2）转换到数值类型时，数值大于type_name可表示的范围时，则会溢出，但不会报错。
-        3）转换到字符串类型时，如果转换后长度超过type_name的长度，则会截断，但不会报错。
+        3）转换到字符串类型时，如果转换后长度超过type_name中指定的长度，则会截断，但不会报错。
 
 #### TO_ISO8601
 
@@ -421,7 +421,7 @@ SELECT CAST(expression AS type_name) FROM { tb_name | stb_name } [WHERE clause]
 SELECT TO_ISO8601(ts[, timezone]) FROM { tb_name | stb_name } [WHERE clause];
 ```
 
-**功能说明**：将 UNIX 时间戳转换成为 ISO8601 标准的日期时间格式，并附加时区信息。timezone 参数允许用户为输出结果指定附带任意时区信息。如果 timezone 参数省略，输出结果附带当前客户端的系统时区信息。
+**功能说明**：将 UNIX 时间戳转换成为 ISO8601 标准的日期时间格式，并附加时区信息。timezone 参数允许用户为输出结果指定附带任意时区信息。如果 timezone 参数省略，输出结果则附带当前客户端的系统时区信息。
 
 **返回结果数据类型**：VARCHAR 类型。
 
@@ -435,7 +435,7 @@ SELECT TO_ISO8601(ts[, timezone]) FROM { tb_name | stb_name } [WHERE clause];
 
 - timezone 参数允许输入的时区格式为: [z/Z, +/-hhmm, +/-hh, +/-hh:mm]。例如，TO_ISO8601(1, "+00:00")。
 - 如果输入是表示 UNIX 时间戳的整形，返回格式精度由时间戳的位数决定; 
-- 如果输入是 TIMSTAMP 类型的列，返回格式的时间戳精度与当前 DATABASE 设置的时间精度一致。
+- 如果输入是 TIMESTAMP 类型的列，返回格式的时间戳精度与当前 DATABASE 设置的时间精度一致。
 
 
 #### TO_JSON
@@ -516,7 +516,7 @@ SELECT TIMEDIFF(ts | datetime_string1, ts | datetime_string2 [, time_unit]) FROM
 
 **功能说明**：计算两个时间戳之间的差值，并近似到时间单位 time_unit 指定的精度。
 
-**返回结果数据类型**：BIGINT。输入包含不符合时间日期格式字符串则返回 NULL。
+**返回结果数据类型**：BIGINT。
 
 **应用字段**：表示 UNIX 时间戳的 BIGINT, TIMESTAMP 类型，或符合日期时间格式的 VARCHAR, NCHAR 类型。
 
@@ -528,6 +528,7 @@ SELECT TIMEDIFF(ts | datetime_string1, ts | datetime_string2 [, time_unit]) FROM
 - 支持的时间单位 time_unit 如下：
           1b(纳秒), 1u(微秒)，1a(毫秒)，1s(秒)，1m(分)，1h(小时)，1d(天), 1w(周)。
 - 如果时间单位 time_unit 未指定， 返回的时间差值精度与当前 DATABASE 设置的时间精度一致。
+- 输入包含不符合时间日期格式的字符串则返回 NULL。
 
 
 #### TIMETRUNCATE
@@ -548,6 +549,7 @@ SELECT TIMETRUNCATE(ts | datetime_string , time_unit) FROM { tb_name | stb_name
 - 支持的时间单位 time_unit 如下：
           1b(纳秒), 1u(微秒)，1a(毫秒)，1s(秒)，1m(分)，1h(小时)，1d(天), 1w(周)。
 - 返回的时间戳精度与当前 DATABASE 设置的时间精度一致。
+- 输入包含不符合时间日期格式的字符串则返回 NULL。
 
 
 #### TIMEZONE
@@ -659,8 +661,6 @@ SELECT ELAPSED(ts_primary_key [, time_unit]) FROM { tb_name | stb_name } [WHERE
 
 **适用数据类型**：TIMESTAMP。
 
-**支持的版本**：2.6.0.0 及以后的版本。
-
 **适用于**: 表，超级表，嵌套查询的外层查询
 
 **说明**：
@@ -767,19 +767,19 @@ SELECT HISTOGRAM(field_name，bin_type, bin_description, normalized) FROM tb_nam
 **适用于**: 表和超级表。
 
 **详细说明**：
-1. bin_type 用户指定的分桶类型, 有效输入类型为"user_input“, ”linear_bin", "log_bin"。
-2. bin_description 描述如何生成分桶区间，针对三种桶类型，分别为以下描述格式(均为 JSON 格式字符串)：       
+- bin_type 用户指定的分桶类型, 有效输入类型为"user_input“, ”linear_bin", "log_bin"。
+- bin_description 描述如何生成分桶区间，针对三种桶类型，分别为以下描述格式(均为 JSON 格式字符串)：       
     - "user_input": "[1, 3, 5, 7]" 
        用户指定 bin 的具体数值。
        
     - "linear_bin": "{"start": 0.0, "width": 5.0, "count": 5, "infinity": true}"
-       "start" 表示数据起始点，"width" 表示每次 bin 偏移量, "count" 为 bin 的总数，"infinity" 表示是否添加（-inf, inf）作为区间起点跟终点，
+       "start" 表示数据起始点，"width" 表示每次 bin 偏移量, "count" 为 bin 的总数，"infinity" 表示是否添加（-inf, inf）作为区间起点和终点，
        生成区间为[-inf, 0.0, 5.0, 10.0, 15.0, 20.0, +inf]。
  
     - "log_bin": "{"start":1.0, "factor": 2.0, "count": 5, "infinity": true}"
-       "start" 表示数据起始点，"factor" 表示按指数递增的因子，"count" 为 bin 的总数，"infinity" 表示是否添加（-inf, inf）作为区间起点跟终点，
+       "start" 表示数据起始点，"factor" 表示按指数递增的因子，"count" 为 bin 的总数，"infinity" 表示是否添加（-inf, inf）作为区间起点和终点，
        生成区间为[-inf, 1.0, 2.0, 4.0, 8.0, 16.0, +inf]。
-3. normalized 是否将返回结果归一化到 0~1 之间 。有效输入为 0 和 1。
+- normalized 是否将返回结果归一化到 0~1 之间 。有效输入为 0 和 1。
 
 
 ### PERCENTILE
@@ -958,20 +958,20 @@ SELECT MODE(field_name) FROM tb_name [WHERE clause];
 SELECT SAMPLE(field_name, K) FROM { tb_name | stb_name } [WHERE clause]
 ```
 
-  **功能说明**： 获取数据的 k 个采样值。参数 k 的合法输入范围是 1≤ k ≤ 1000。
+**功能说明**： 获取数据的 k 个采样值。参数 k 的合法输入范围是 1≤ k ≤ 1000。
 
-  **返回结果类型**： 同原始数据类型， 返回结果中带有该行记录的时间戳。
+**返回结果类型**： 同原始数据类型， 返回结果中带有该行记录的时间戳。
 
-  **适用数据类型**： 在超级表查询中使用时，不能应用在标签之上。
+**适用数据类型**： 在超级表查询中使用时，不能应用在标签之上。
 
-  **嵌套子查询支持**： 适用于内层查询和外层查询。
+**嵌套子查询支持**： 适用于内层查询和外层查询。
 
-  **适用于**：表和超级表。
+**适用于**：表和超级表。
 
-  **使用说明**： 
+**使用说明**： 
 
-  - 不能参与表达式计算；该函数可以应用在普通表和超级表上；
-  - 使用在超级表上的时候，需要搭配 PARTITION by tbname 使用，将结果强制规约到单个时间线。
+- 不能参与表达式计算；该函数可以应用在普通表和超级表上；
+- 使用在超级表上的时候，需要搭配 PARTITION by tbname 使用，将结果强制规约到单个时间线。
 
 
 ### TAIL
@@ -1048,9 +1048,9 @@ SELECT CSUM(field_name) FROM { tb_name | stb_name } [WHERE clause]
 
 **使用说明**： 
   
-  - 不支持 +、-、*、/ 运算，如 csum(col1) + csum(col2)。
-  - 只能与聚合（Aggregation）函数一起使用。 该函数可以应用在普通表和超级表上。 
-  - 使用在超级表上的时候，需要搭配 PARTITION BY tbname使用，将结果强制规约到单个时间线。
+- 不支持 +、-、*、/ 运算，如 csum(col1) + csum(col2)。
+- 只能与聚合（Aggregation）函数一起使用。 该函数可以应用在普通表和超级表上。 
+- 使用在超级表上的时候，需要搭配 PARTITION BY tbname使用，将结果强制规约到单个时间线。
 
 
 ### DERIVATIVE
@@ -1069,8 +1069,8 @@ SELECT DERIVATIVE(field_name, time_interval, ignore_negative) FROM tb_name [WHER
 
 **使用说明**: 
   
-  - DERIVATIVE 函数可以在由 PARTITION BY 划分出单独时间线的情况下用于超级表（也即 PARTITION BY tbname）。
-  - 可以与选择相关联的列一起使用。 例如: select \_rowts, DERIVATIVE() from。
+- DERIVATIVE 函数可以在由 PARTITION BY 划分出单独时间线的情况下用于超级表（也即 PARTITION BY tbname）。
+- 可以与选择相关联的列一起使用。 例如: select \_rowts, DERIVATIVE() from。
 
 ### DIFF
 
@@ -1088,8 +1088,8 @@ SELECT {DIFF(field_name, ignore_negative) | DIFF(field_name)} FROM tb_name [WHER
 
 **使用说明**: 
 
-  - 输出结果行数是范围内总行数减一，第一行没有结果输出。
-  - 可以与选择相关联的列一起使用。 例如: select \_rowts, DIFF() from。
+- 输出结果行数是范围内总行数减一，第一行没有结果输出。
+- 可以与选择相关联的列一起使用。 例如: select \_rowts, DIFF() from。
 
 
 ### IRATE
@@ -1113,21 +1113,21 @@ SELECT IRATE(field_name) FROM tb_name WHERE clause;
 SELECT MAVG(field_name, K) FROM { tb_name | stb_name } [WHERE clause]
 ```
 
-  **功能说明**： 计算连续 k 个值的移动平均数（moving average）。如果输入行数小于 k，则无结果输出。参数 k 的合法输入范围是 1≤ k ≤ 1000。
+**功能说明**： 计算连续 k 个值的移动平均数（moving average）。如果输入行数小于 k，则无结果输出。参数 k 的合法输入范围是 1≤ k ≤ 1000。
 
-  **返回结果类型**： DOUBLE。
+**返回结果类型**： DOUBLE。
 
-  **适用数据类型**： 数值类型。
+**适用数据类型**： 数值类型。
 
-  **嵌套子查询支持**： 适用于内层查询和外层查询。
+**嵌套子查询支持**： 适用于内层查询和外层查询。
 
-  **适用于**：表和超级表。
+**适用于**：表和超级表。
 
-  **使用说明**： 
+**使用说明**： 
   
-  - 不支持 +、-、*、/ 运算，如 mavg(col1, k1) + mavg(col2, k1); 
-  - 只能与普通列，选择（Selection）、投影（Projection）函数一起使用，不能与聚合（Aggregation）函数一起使用；
-  - 使用在超级表上的时候，需要搭配 PARTITION BY tbname使用，将结果强制规约到单个时间线。
+- 不支持 +、-、*、/ 运算，如 mavg(col1, k1) + mavg(col2, k1); 
+- 只能与普通列，选择（Selection）、投影（Projection）函数一起使用，不能与聚合（Aggregation）函数一起使用；
+- 使用在超级表上的时候，需要搭配 PARTITION BY tbname使用，将结果强制规约到单个时间线。
 
 
 ### STATECOUNT

examples/c/tmq.c

@@ -302,7 +302,7 @@ int32_t create_topic() {
   }
   taos_free_result(pRes);
 
-//  pRes = taos_query(pConn, "create topic topic_ctb_column with meta as database abc1");
+  //  pRes = taos_query(pConn, "create topic topic_ctb_column with meta as database abc1");
   pRes = taos_query(pConn, "create topic topic_ctb_column as select ts, c1, c2, c3 from st1");
   if (taos_errno(pRes) != 0) {
     printf("failed to create topic topic_ctb_column, reason:%s\n", taos_errstr(pRes));

include/common/tglobal.h

@@ -146,6 +146,7 @@ struct SConfig *taosGetCfg();
 void    taosSetAllDebugFlag(int32_t flag);
 void    taosSetDebugFlag(int32_t *pFlagPtr, const char *flagName, int32_t flagVal);
 int32_t taosSetCfg(SConfig *pCfg, char *name);
+void    taosLocalCfgForbiddenToChange(char* name, bool* forbidden);
 
 #ifdef __cplusplus
 }

include/common/ttypes.h

@@ -260,8 +260,7 @@ typedef struct {
     }                                                                                  \
   } while (0)
 
-
-//TODO: use varchar(0) to represent NULL type
+// TODO: use varchar(0) to represent NULL type
 #define IS_VAR_NULL_TYPE(_t, _b) ((_t) == TSDB_DATA_TYPE_VARCHAR && (_b) == 0)
 #define IS_NULL_TYPE(_t)         ((_t) == TSDB_DATA_TYPE_NULL)
 
@@ -329,7 +328,7 @@ typedef struct tDataTypeDescriptor {
   int16_t type;
   int16_t nameLen;
   int32_t bytes;
-  char *  name;
+  char   *name;
   int64_t minValue;
   int64_t maxValue;
   int32_t (*compFunc)(const char *const input, int32_t inputSize, const int32_t nelements, char *const output,

include/libs/nodes/querynodes.h

@@ -418,8 +418,6 @@ void    nodesValueNodeToVariant(const SValueNode* pNode, SVariant* pVal);
 
 char*   nodesGetFillModeString(EFillMode mode);
 int32_t nodesMergeConds(SNode** pDst, SNodeList** pSrc);
-int32_t nodesPartitionCond(SNode** pCondition, SNode** pPrimaryKeyCond, SNode** pTagIndexCond, SNode** pTagCond,
-                           SNode** pOtherCond);
 
 #ifdef __cplusplus
 }

include/libs/scalar/filter.h

@@ -46,6 +46,10 @@ extern int32_t filterFreeNcharColumns(SFilterInfo *pFilterInfo);
 extern void    filterFreeInfo(SFilterInfo *info);
 extern bool    filterRangeExecute(SFilterInfo *info, SColumnDataAgg *pDataStatis, int32_t numOfCols, int32_t numOfRows);
 
+/* condition split interface */
+int32_t filterPartitionCond(SNode **pCondition, SNode **pPrimaryKeyCond, SNode **pTagIndexCond, SNode **pTagCond,
+                            SNode **pOtherCond);
+
 #ifdef __cplusplus
 }
 #endif

include/libs/scalar/scalar.h

@@ -25,7 +25,7 @@ extern "C" {
 
 typedef struct SFilterInfo SFilterInfo;
 
-int32_t scalarGetOperatorResultType(SOperatorNode* pOp);
+int32_t scalarGetOperatorResultType(SOperatorNode *pOp);
 
 /*
 pNode will be freed in API;
@@ -43,7 +43,7 @@ int32_t scalarGetOperatorParamNum(EOperatorType type);
 int32_t scalarGenerateSetFromList(void **data, void *pNode, uint32_t type);
 
 int32_t vectorGetConvertType(int32_t type1, int32_t type2);
-int32_t vectorConvertImpl(const SScalarParam* pIn, SScalarParam* pOut, int32_t* overflow);
+int32_t vectorConvertImpl(const SScalarParam *pIn, SScalarParam *pOut, int32_t *overflow);
 
 /* Math functions */
 int32_t absFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput);
@@ -86,7 +86,7 @@ int32_t nowFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutpu
 int32_t todayFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput);
 int32_t timezoneFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput);
 
-bool getTimePseudoFuncEnv(struct SFunctionNode* pFunc, SFuncExecEnv* pEnv);
+bool getTimePseudoFuncEnv(struct SFunctionNode *pFunc, SFuncExecEnv *pEnv);
 
 int32_t winStartTsFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput);
 int32_t winEndTsFunction(SScalarParam *pInput, int32_t inputNum, SScalarParam *pOutput);

include/os/osString.h

@@ -62,6 +62,8 @@ typedef int32_t TdUcs4;
 int32_t taosUcs4len(TdUcs4 *ucs4);
 int64_t taosStr2int64(const char *str);
 
+void    taosConvInit(void);
+void    taosConvDestroy();
 int32_t taosUcs4ToMbs(TdUcs4 *ucs4, int32_t ucs4_max_len, char *mbs);
 bool    taosMbsToUcs4(const char *mbs, size_t mbs_len, TdUcs4 *ucs4, int32_t ucs4_max_len, int32_t *len);
 int32_t tasoUcs4Compare(TdUcs4 *f1_ucs4, TdUcs4 *f2_ucs4, int32_t bytes);

include/util/types.h

@@ -85,7 +85,7 @@ typedef uint16_t VarDataLenT;  // maxVarDataLen: 32767
 #define varDataVal(v)  ((char *)(v) + VARSTR_HEADER_SIZE)
 #define varDataTLen(v) (sizeof(VarDataLenT) + varDataLen(v))
 
-#define NCHAR_WIDTH_TO_BYTES(n)  ((n) * TSDB_NCHAR_SIZE + VARSTR_HEADER_SIZE)
+#define NCHAR_WIDTH_TO_BYTES(n) ((n)*TSDB_NCHAR_SIZE + VARSTR_HEADER_SIZE)
 
 typedef int32_t VarDataOffsetT;
 

source/client/src/clientEnv.c

@@ -361,6 +361,8 @@ void taos_init_imp(void) {
 
   initQueryModuleMsgHandle();
 
+  taosConvInit();
+  
   rpcInit();
 
   SCatalogCfg cfg = {.maxDBCacheNum = 100, .maxTblCacheNum = 100};

source/client/src/clientHb.c

@@ -353,6 +353,7 @@ int32_t hbBuildQueryDesc(SQueryHbReqBasic *hbBasic, STscObj *pObj) {
         desc.subDesc = NULL;
         desc.subPlanNum = 0;
       }
+      desc.subPlanNum = taosArrayGetSize(desc.subDesc);
       ASSERT(desc.subPlanNum == taosArrayGetSize(desc.subDesc));
     } else {
       desc.subDesc = NULL;

source/client/src/clientMain.c

@@ -75,6 +75,8 @@ void taos_cleanup(void) {
 
   cleanupTaskQueue();
 
+  taosConvDestroy();
+  
   tscInfo("all local resources released");
   taosCleanupCfg();
   taosCloseLog();

source/common/src/tglobal.c

@@ -49,7 +49,7 @@ int32_t tsNumOfShmThreads = 1;
 // queue & threads
 int32_t tsNumOfRpcThreads = 1;
 int32_t tsNumOfCommitThreads = 2;
-int32_t tsNumOfTaskQueueThreads = 1;
+int32_t tsNumOfTaskQueueThreads = 4;
 int32_t tsNumOfMnodeQueryThreads = 4;
 int32_t tsNumOfMnodeFetchThreads = 1;
 int32_t tsNumOfMnodeReadThreads = 1;
@@ -317,9 +317,9 @@ static int32_t taosAddClientCfg(SConfig *pCfg) {
   if (cfgAddString(pCfg, "smlTagName", tsSmlTagName, 1) != 0) return -1;
   if (cfgAddBool(pCfg, "smlDataFormat", tsSmlDataFormat, 1) != 0) return -1;
 
-  tsNumOfTaskQueueThreads = tsNumOfCores / 4;
-  tsNumOfTaskQueueThreads = TRANGE(tsNumOfTaskQueueThreads, 1, 2);
-  if (cfgAddInt32(pCfg, "numOfTaskQueueThreads", tsNumOfTaskQueueThreads, 1, 1024, 0) != 0) return -1;
+  tsNumOfTaskQueueThreads = tsNumOfCores / 2;
+  tsNumOfTaskQueueThreads = TMAX(tsNumOfTaskQueueThreads, 4);
+  if (cfgAddInt32(pCfg, "numOfTaskQueueThreads", tsNumOfTaskQueueThreads, 4, 1024, 0) != 0) return -1;
 
   return 0;
 }
@@ -594,6 +594,20 @@ static int32_t taosSetServerCfg(SConfig *pCfg) {
   return 0;
 }
 
+void taosLocalCfgForbiddenToChange(char* name, bool* forbidden) {
+  int32_t len = strlen(name);
+  char    lowcaseName[CFG_NAME_MAX_LEN + 1] = {0};
+  strntolower(lowcaseName, name, TMIN(CFG_NAME_MAX_LEN, len));
+
+  if (strcasecmp("charset", name) == 0) {
+    *forbidden = true;
+    return;
+  }
+
+  *forbidden = false;
+}
+
+
 int32_t taosSetCfg(SConfig *pCfg, char *name) {
   int32_t len = strlen(name);
   char    lowcaseName[CFG_NAME_MAX_LEN + 1] = {0};

source/dnode/mgmt/exe/dmMain.c

@@ -219,6 +219,8 @@ int mainWindows(int argc,char** argv) {
     return -1;
   }
   
+  taosConvInit();
+
   if (global.dumpConfig) {
     dmDumpCfg();
     taosCleanupCfg();

source/dnode/mgmt/node_mgmt/src/dmMgmt.c

@@ -215,6 +215,7 @@ void dmCleanupDnode(SDnode *pDnode) {
   dmClearVars(pDnode);
   rpcCleanup();
   indexCleanup();
+  taosConvDestroy();
   dDebug("dnode is closed, ptr:%p", pDnode);
 }
 

source/dnode/vnode/inc/vnode.h

@@ -89,13 +89,13 @@ void        metaReaderClear(SMetaReader *pReader);
 int32_t     metaGetTableEntryByUid(SMetaReader *pReader, tb_uid_t uid);
 int32_t     metaReadNext(SMetaReader *pReader);
 const void *metaGetTableTagVal(SMetaEntry *pEntry, int16_t type, STagVal *tagVal);
-int         metaGetTableNameByUid(void* meta, uint64_t uid, char* tbName);
+int         metaGetTableNameByUid(void *meta, uint64_t uid, char *tbName);
 
 typedef struct SMetaFltParam {
   tb_uid_t suid;
   int16_t  cid;
   int16_t  type;
-  char    *val;
+  void    *val;
   bool     reverse;
   int (*filterFunc)(void *a, void *b, int16_t type);
 

source/dnode/vnode/src/inc/tsdb.h

@@ -90,6 +90,9 @@ int32_t tsdbRowCmprFn(const void *p1, const void *p2);
 void     tRowIterInit(SRowIter *pIter, TSDBROW *pRow, STSchema *pTSchema);
 SColVal *tRowIterNext(SRowIter *pIter);
 // SRowMerger
+int32_t tRowMergerInit2(SRowMerger *pMerger, STSchema *pResTSchema, TSDBROW *pRow, STSchema *pTSchema);
+int32_t tRowMergerAdd(SRowMerger *pMerger, TSDBROW *pRow, STSchema *pTSchema);
+
 int32_t tRowMergerInit(SRowMerger *pMerger, TSDBROW *pRow, STSchema *pTSchema);
 void    tRowMergerClear(SRowMerger *pMerger);
 int32_t tRowMerge(SRowMerger *pMerger, TSDBROW *pRow);

source/dnode/vnode/src/inc/vnodeInt.h

@@ -135,6 +135,7 @@ int32_t     tsdbInsertTableData(STsdb* pTsdb, int64_t version, SSubmitMsgIter* p
 int32_t     tsdbDeleteTableData(STsdb* pTsdb, int64_t version, tb_uid_t suid, tb_uid_t uid, TSKEY sKey, TSKEY eKey);
 STsdbReader tsdbQueryCacheLastT(STsdb* tsdb, SQueryTableDataCond* pCond, STableListInfo* tableList, uint64_t qId,
                                 void* pMemRef);
+int32_t     tsdbSetKeepCfg(STsdb* pTsdb, STsdbCfg* pCfg);
 
 // tq
 int     tqInit();

source/dnode/vnode/src/meta/metaOpen.c

@@ -315,7 +315,8 @@ static int tagIdxKeyCmpr(const void *pKey1, int kLen1, const void *pKey2, int kL
     return 1;
   } else if (!pTagIdxKey1->isNull && !pTagIdxKey2->isNull) {
     // all not NULL, compr tag vals
-    c = doCompare(pTagIdxKey1->data, pTagIdxKey2->data, pTagIdxKey1->type, 0);
+    __compar_fn_t func = getComparFunc(pTagIdxKey1->type, 0);
+    c = func(pTagIdxKey1->data, pTagIdxKey2->data);
     if (c) return c;
   }
 

source/dnode/vnode/src/meta/metaQuery.c

@@ -98,9 +98,9 @@ tb_uid_t metaGetTableEntryUidByName(SMeta *pMeta, const char *name) {
   return uid;
 }
 
-int metaGetTableNameByUid(void* meta, uint64_t uid, char* tbName) {
+int metaGetTableNameByUid(void *meta, uint64_t uid, char *tbName) {
   SMetaReader mr = {0};
-  metaReaderInit(&mr, (SMeta*)meta, 0);
+  metaReaderInit(&mr, (SMeta *)meta, 0);
   metaGetTableEntryByUid(&mr, uid);
 
   STR_TO_VARSTR(tbName, mr.me.name);

source/dnode/vnode/src/meta/metaTable.c

@@ -793,9 +793,6 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
     goto _err;
   }
 
-  if (iCol == 0) {
-    // TODO : need to update tag index
-  }
   ctbEntry.version = version;
   if (pTagSchema->nCols == 1 && pTagSchema->pSchema[0].type == TSDB_DATA_TYPE_JSON) {
     ctbEntry.ctbEntry.pTags = taosMemoryMalloc(pAlterTbReq->nTagVal);
@@ -849,6 +846,10 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
   // save to uid.idx
   tdbTbUpsert(pMeta->pUidIdx, &ctbEntry.uid, sizeof(tb_uid_t), &version, sizeof(version), &pMeta->txn);
 
+  if (iCol == 0) {
+    metaUpdateTagIdx(pMeta, &ctbEntry);
+  }
+
   tDecoderClear(&dc1);
   tDecoderClear(&dc2);
   if (ctbEntry.ctbEntry.pTags) taosMemoryFree((void *)ctbEntry.ctbEntry.pTags);
@@ -1131,7 +1132,7 @@ static int metaUpdateTagIdx(SMeta *pMeta, const SMetaEntry *pCtbEntry) {
                           pCtbEntry->uid, &pTagIdxKey, &nTagIdxKey) < 0) {
     return -1;
   }
-  tdbTbInsert(pMeta->pTagIdx, pTagIdxKey, nTagIdxKey, NULL, 0, &pMeta->txn);
+  tdbTbUpsert(pMeta->pTagIdx, pTagIdxKey, nTagIdxKey, NULL, 0, &pMeta->txn);
   metaDestroyTagIdxKey(pTagIdxKey);
   tDecoderClear(&dc);
   tdbFree(pData);

source/dnode/vnode/src/tq/tq.c

@@ -684,6 +684,9 @@ int32_t tqProcessTaskDeployReq(STQ* pTq, char* msg, int32_t msgLen) {
 
   taosHashPut(pTq->pStreamTasks, &pTask->taskId, sizeof(int32_t), &pTask, sizeof(void*));
 
+  /*SMeta* pMeta = pTq->pVnode->pMeta;*/
+  /*tdbTbUpsert(pMeta->pTaskIdx, &pTask->taskId, sizeof(int32_t), msg, msgLen, &pMeta->txn);*/
+
   return 0;
 FAIL:
   if (pTask->inputQueue) streamQueueClose(pTask->inputQueue);

source/dnode/vnode/src/tsdb/tsdbCommit.c

@@ -311,6 +311,8 @@ static int32_t tsdbCommitFileDataStart(SCommitter *pCommitter) {
 
     tfsAllocDisk(pTsdb->pVnode->pTfs, 0, &did);
 
+    tfsMkdirRecurAt(pTsdb->pVnode->pTfs, pTsdb->path, did);
+
     wSet.diskId = did;
     wSet.fid = pCommitter->commitFid;
     fHead = (SHeadFile){.commitID = pCommitter->commitID, .offset = 0, .size = 0};

source/dnode/vnode/src/tsdb/tsdbOpen.c

@@ -15,11 +15,8 @@
 
 #include "tsdb.h"
 
-static int tsdbSetKeepCfg(STsdbKeepCfg *pKeepCfg, STsdbCfg *pCfg);
-
-// implementation
-
-static int tsdbSetKeepCfg(STsdbKeepCfg *pKeepCfg, STsdbCfg *pCfg) {
+int32_t tsdbSetKeepCfg(STsdb *pTsdb, STsdbCfg *pCfg) {
+  STsdbKeepCfg *pKeepCfg = &pTsdb->keepCfg;
   pKeepCfg->precision = pCfg->precision;
   pKeepCfg->days = pCfg->days;
   pKeepCfg->keep0 = pCfg->keep0;
@@ -56,7 +53,7 @@ int tsdbOpen(SVnode *pVnode, STsdb **ppTsdb, const char *dir, STsdbKeepCfg *pKee
   pTsdb->pVnode = pVnode;
   taosThreadRwlockInit(&pTsdb->rwLock, NULL);
   if (!pKeepCfg) {
-    tsdbSetKeepCfg(&pTsdb->keepCfg, &pVnode->config.tsdbCfg);
+    tsdbSetKeepCfg(pTsdb, &pVnode->config.tsdbCfg);
   } else {
     memcpy(&pTsdb->keepCfg, pKeepCfg, sizeof(STsdbKeepCfg));
   }

source/dnode/vnode/src/tsdb/tsdbRead.c

@@ -143,7 +143,7 @@ static int      buildDataBlockFromBufImpl(STableBlockScanInfo* pBlockScanInfo, i
 static TSDBROW* getValidRow(SIterInfo* pIter, const SArray* pDelList, STsdbReader* pReader);
 static int32_t  doMergeRowsInFileBlocks(SBlockData* pBlockData, STableBlockScanInfo* pScanInfo, STsdbReader* pReader,
                                         SRowMerger* pMerger);
-static int32_t  doMergeRowsInBuf(SIterInfo* pIter, int64_t ts, SArray* pDelList, SRowMerger* pMerger,
+static int32_t  doMergeRowsInBuf(SIterInfo* pIter, uint64_t uid, int64_t ts, SArray* pDelList, SRowMerger* pMerger,
                                  STsdbReader* pReader);
 static int32_t  doAppendRowFromTSRow(SSDataBlock* pBlock, STsdbReader* pReader, STSRow* pTSRow);
 static int32_t  doAppendRowFromBlock(SSDataBlock* pResBlock, STsdbReader* pReader, SBlockData* pBlockData, int32_t rowIndex);
@@ -212,8 +212,8 @@ static SHashObj* createDataBlockScanInfo(STsdbReader* pTsdbReader, const STableK
               pTsdbReader->idStr);
   }
 
-  tsdbDebug("%p create %d tables scan-info, size:%.2f Kb, %s", pTsdbReader, numOfTables, (sizeof(STableBlockScanInfo)*numOfTables)/1024.0,
-      pTsdbReader->idStr);
+  tsdbDebug("%p create %d tables scan-info, size:%.2f Kb, %s", pTsdbReader, numOfTables,
+            (sizeof(STableBlockScanInfo) * numOfTables) / 1024.0, pTsdbReader->idStr);
 
   return pTableMap;
 }
@@ -397,7 +397,8 @@ static SSDataBlock* createResBlock(SQueryTableDataCond* pCond, int32_t capacity)
   return pResBlock;
 }
 
-static int32_t tsdbReaderCreate(SVnode* pVnode, SQueryTableDataCond* pCond, STsdbReader** ppReader, int32_t capacity, const char* idstr) {
+static int32_t tsdbReaderCreate(SVnode* pVnode, SQueryTableDataCond* pCond, STsdbReader** ppReader, int32_t capacity,
+                                const char* idstr) {
   int32_t      code = 0;
   int8_t       level = 0;
   STsdbReader* pReader = (STsdbReader*)taosMemoryCalloc(1, sizeof(*pReader));
@@ -577,7 +578,7 @@ static int32_t doLoadBlockIndex(STsdbReader* pReader, SDataFReader* pFileReader,
 
   int64_t et2 = taosGetTimestampUs();
   tsdbDebug("load block index for %d tables completed, elapsed time:%.2f ms, set blockIdx:%.2f ms, size:%.2f Kb %s",
-            (int32_t)num, (et1 - st)/1000.0, (et2-et1)/1000.0, num * sizeof(SBlockIdx)/1024.0, pReader->idStr);
+            (int32_t)num, (et1 - st) / 1000.0, (et2 - et1) / 1000.0, num * sizeof(SBlockIdx) / 1024.0, pReader->idStr);
 
   pReader->cost.headFileLoadTime += (et1 - st) / 1000.0;
 
@@ -642,9 +643,9 @@ static int32_t doLoadFileBlock(STsdbReader* pReader, SArray* pIndexList, uint32_
     }
   }
 
-  double el = (taosGetTimestampUs() - st)/1000.0;
+  double el = (taosGetTimestampUs() - st) / 1000.0;
   tsdbDebug("load block of %d tables completed, blocks:%d in %d tables, size:%.2f Kb, elapsed time:%.2f ms %s",
-      numOfTables, *numOfBlocks, *numOfValidTables, size/1000.0, el, pReader->idStr);
+            numOfTables, *numOfBlocks, *numOfValidTables, size / 1000.0, el, pReader->idStr);
 
   pReader->cost.numOfBlocks += (*numOfBlocks);
   pReader->cost.headFileLoadTime += el;
@@ -680,9 +681,7 @@ static SFileDataBlockInfo* getCurrentBlockInfo(SDataBlockIter* pBlockIter) {
   return pFBlockInfo;
 }
 
-static SBlock* getCurrentBlock(SDataBlockIter* pBlockIter) {
-  return &pBlockIter->block;
-}
+static SBlock* getCurrentBlock(SDataBlockIter* pBlockIter) { return &pBlockIter->block; }
 
 static int32_t copyBlockDataToSDataBlock(STsdbReader* pReader, STableBlockScanInfo* pBlockScanInfo) {
   SReaderStatus*  pStatus = &pReader->status;
@@ -781,8 +780,8 @@ static int32_t doLoadFileBlockData(STsdbReader* pReader, SDataBlockIter* pBlockI
   SFileBlockDumpInfo* pDumpInfo = &pReader->status.fBlockDumpInfo;
 
   SBlockIdx blockIdx = {.suid = pReader->suid, .uid = pBlockScanInfo->uid};
-  int32_t code = tsdbReadColData(pReader->pFileReader, &blockIdx, pBlock, pSupInfo->colIds, numOfCols,
-                                 pBlockData, NULL, NULL);
+  int32_t   code =
+      tsdbReadColData(pReader->pFileReader, &blockIdx, pBlock, pSupInfo->colIds, numOfCols, pBlockData, NULL, NULL);
   if (code != TSDB_CODE_SUCCESS) {
     goto _error;
   }
@@ -937,8 +936,8 @@ static int32_t initBlockIterator(STsdbReader* pReader, SDataBlockIter* pBlockIte
     }
 
     int64_t et = taosGetTimestampUs();
-    tsdbDebug("%p create blocks info struct completed for one table, %d blocks not sorted, elapsed time:%.2f ms %s", pReader, cnt,
-              (et - st)/1000.0, pReader->idStr);
+    tsdbDebug("%p create blocks info struct completed for one table, %d blocks not sorted, elapsed time:%.2f ms %s",
+              pReader, cnt, (et - st) / 1000.0, pReader->idStr);
 
     pBlockIter->index = asc ? 0 : (numOfBlocks - 1);
     cleanupBlockOrderSupporter(&sup);
@@ -976,7 +975,8 @@ static int32_t initBlockIterator(STsdbReader* pReader, SDataBlockIter* pBlockIte
   }
 
   int64_t et = taosGetTimestampUs();
-  tsdbDebug("%p %d data blocks access order completed, elapsed time:%.2f ms %s", pReader, cnt, (et-st)/1000.0, pReader->idStr);
+  tsdbDebug("%p %d data blocks access order completed, elapsed time:%.2f ms %s", pReader, cnt, (et - st) / 1000.0,
+            pReader->idStr);
   cleanupBlockOrderSupporter(&sup);
   taosMemoryFree(pTree);
 
@@ -1024,7 +1024,7 @@ static SBlock* getNeighborBlockOfSameTable(SFileDataBlockInfo* pFBlockInfo, STab
   int32_t step = asc ? 1 : -1;
   *nextIndex = pFBlockInfo->tbBlockIdx + step;
 
-  SBlock *pBlock = taosMemoryCalloc(1, sizeof(SBlock));
+  SBlock*  pBlock = taosMemoryCalloc(1, sizeof(SBlock));
   int32_t* indexInMapdata = taosArrayGet(pTableBlockScanInfo->pBlockList, *nextIndex);
 
   tMapDataGetItemByIdx(&pTableBlockScanInfo->mapData, *indexInMapdata, pBlock, tGetBlock);
@@ -1196,10 +1196,10 @@ static int32_t buildDataBlockFromBuf(STsdbReader* pReader, STableBlockScanInfo*
   setComposedBlockFlag(pReader, true);
 
   double elapsedTime = (taosGetTimestampUs() - st) / 1000.0;
-  tsdbDebug(
-      "%p build data block from cache completed, elapsed time:%.2f ms, numOfRows:%d, brange: %" PRId64
+  tsdbDebug("%p build data block from cache completed, elapsed time:%.2f ms, numOfRows:%d, brange: %" PRId64
             " - %" PRId64 " %s",
-      pReader, elapsedTime, pBlock->info.rows, pBlock->info.window.skey, pBlock->info.window.ekey, pReader->idStr);
+            pReader, elapsedTime, pBlock->info.rows, pBlock->info.window.skey, pBlock->info.window.ekey,
+            pReader->idStr);
 
   pReader->cost.buildmemBlock += elapsedTime;
   return code;
@@ -1230,7 +1230,7 @@ static int32_t doMergeBufAndFileRows(STsdbReader* pReader, STableBlockScanInfo*
       doMergeRowsInFileBlocks(pBlockData, pBlockScanInfo, pReader, &merge);
 
       tRowMerge(&merge, pRow);
-      doMergeRowsInBuf(pIter, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
+      doMergeRowsInBuf(pIter, pBlockScanInfo->uid, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
 
       tRowMergerGetRow(&merge, &pTSRow);
     }
@@ -1246,7 +1246,7 @@ static int32_t doMergeBufAndFileRows(STsdbReader* pReader, STableBlockScanInfo*
       updateSchema(pRow, pBlockScanInfo->uid, pReader);
 
       tRowMergerInit(&merge, pRow, pReader->pSchema);
-      doMergeRowsInBuf(pIter, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
+      doMergeRowsInBuf(pIter, pBlockScanInfo->uid, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
 
       tRowMerge(&merge, &fRow);
       doMergeRowsInFileBlocks(pBlockData, pBlockScanInfo, pReader, &merge);
@@ -1291,12 +1291,12 @@ static int32_t doMergeThreeLevelRows(STsdbReader* pReader, STableBlockScanInfo*
 
       if (ik.ts == key) {
         tRowMerge(&merge, piRow);
-        doMergeRowsInBuf(&pBlockScanInfo->iiter, key, pBlockScanInfo->delSkyline, &merge, pReader);
+        doMergeRowsInBuf(&pBlockScanInfo->iiter, uid, key, pBlockScanInfo->delSkyline, &merge, pReader);
       }
 
       if (k.ts == key) {
         tRowMerge(&merge, pRow);
-        doMergeRowsInBuf(&pBlockScanInfo->iter, key, pBlockScanInfo->delSkyline, &merge, pReader);
+        doMergeRowsInBuf(&pBlockScanInfo->iter, uid, key, pBlockScanInfo->delSkyline, &merge, pReader);
       }
 
       tRowMergerGetRow(&merge, &pTSRow);
@@ -1336,11 +1336,11 @@ static int32_t doMergeThreeLevelRows(STsdbReader* pReader, STableBlockScanInfo*
       updateSchema(pRow, uid, pReader);
 
       tRowMergerInit(&merge, pRow, pReader->pSchema);
-      doMergeRowsInBuf(&pBlockScanInfo->iter, key, pBlockScanInfo->delSkyline, &merge, pReader);
+      doMergeRowsInBuf(&pBlockScanInfo->iter, uid, key, pBlockScanInfo->delSkyline, &merge, pReader);
 
       if (ik.ts == k.ts) {
         tRowMerge(&merge, piRow);
-        doMergeRowsInBuf(&pBlockScanInfo->iiter, key, pBlockScanInfo->delSkyline, &merge, pReader);
+        doMergeRowsInBuf(&pBlockScanInfo->iiter, uid, key, pBlockScanInfo->delSkyline, &merge, pReader);
       }
 
       if (k.ts == key) {
@@ -1514,9 +1514,10 @@ static int32_t buildComposedDataBlock(STsdbReader* pReader, STableBlockScanInfo*
   setComposedBlockFlag(pReader, true);
   int64_t et = taosGetTimestampUs();
 
-  tsdbDebug("%p uid:%" PRIu64 ", composed data block created, brange:%" PRIu64 "-%" PRIu64 " rows:%d, elapsed time:%.2f ms %s", pReader,
-            pBlockScanInfo->uid, pResBlock->info.window.skey, pResBlock->info.window.ekey, pResBlock->info.rows,
-            (et - st)/1000.0, pReader->idStr);
+  tsdbDebug("%p uid:%" PRIu64 ", composed data block created, brange:%" PRIu64 "-%" PRIu64
+            " rows:%d, elapsed time:%.2f ms %s",
+            pReader, pBlockScanInfo->uid, pResBlock->info.window.skey, pResBlock->info.window.ekey,
+            pResBlock->info.rows, (et - st) / 1000.0, pReader->idStr);
 
   return TSDB_CODE_SUCCESS;
 }
@@ -2111,7 +2112,8 @@ TSDBROW* getValidRow(SIterInfo* pIter, const SArray* pDelList, STsdbReader* pRea
   }
 }
 
-int32_t doMergeRowsInBuf(SIterInfo* pIter, int64_t ts, SArray* pDelList, SRowMerger* pMerger, STsdbReader* pReader) {
+int32_t doMergeRowsInBuf(SIterInfo* pIter, uint64_t uid, int64_t ts, SArray* pDelList, SRowMerger* pMerger,
+                         STsdbReader* pReader) {
   while (1) {
     pIter->hasVal = tsdbTbDataIterNext(pIter->iter);
     if (!pIter->hasVal) {
@@ -2130,7 +2132,19 @@ int32_t doMergeRowsInBuf(SIterInfo* pIter, int64_t ts, SArray* pDelList, SRowMer
       break;
     }
 
-    tRowMerge(pMerger, pRow);
+    int32_t   sversion = TSDBROW_SVERSION(pRow);
+    STSchema* pTSchema = NULL;
+    if (sversion != pReader->pSchema->version) {
+      metaGetTbTSchemaEx(pReader->pTsdb->pVnode->pMeta, pReader->suid, uid, sversion, &pTSchema);
+    } else {
+      pTSchema = pReader->pSchema;
+    }
+
+    tRowMergerAdd(pMerger, pRow, pTSchema);
+
+    if (sversion != pReader->pSchema->version) {
+      taosMemoryFree(pTSchema);
+    }
   }
 
   return TSDB_CODE_SUCCESS;
@@ -2254,12 +2268,23 @@ void doMergeMultiRows(TSDBROW* pRow, uint64_t uid, SIterInfo* pIter, SArray* pDe
   SRowMerger merge = {0};
 
   TSDBKEY k = TSDBROW_KEY(pRow);
-  updateSchema(pRow, uid, pReader);
+  // updateSchema(pRow, uid, pReader);
+  int32_t   sversion = TSDBROW_SVERSION(pRow);
+  STSchema* pTSchema = NULL;
+  if (sversion != pReader->pSchema->version) {
+    metaGetTbTSchemaEx(pReader->pTsdb->pVnode->pMeta, pReader->suid, uid, sversion, &pTSchema);
+  } else {
+    pTSchema = pReader->pSchema;
+  }
 
-  tRowMergerInit(&merge, pRow, pReader->pSchema);
-  doMergeRowsInBuf(pIter, k.ts, pDelList, &merge, pReader);
+  tRowMergerInit2(&merge, pReader->pSchema, pRow, pTSchema);
+  doMergeRowsInBuf(pIter, uid, k.ts, pDelList, &merge, pReader);
   tRowMergerGetRow(&merge, pTSRow);
   tRowMergerClear(&merge);
+
+  if (sversion != pReader->pSchema->version) {
+    taosMemoryFree(pTSchema);
+  }
 }
 
 void doMergeMemIMemRows(TSDBROW* pRow, TSDBROW* piRow, STableBlockScanInfo* pBlockScanInfo, STsdbReader* pReader,
@@ -2273,18 +2298,18 @@ void doMergeMemIMemRows(TSDBROW* pRow, TSDBROW* piRow, STableBlockScanInfo* pBlo
     updateSchema(piRow, pBlockScanInfo->uid, pReader);
 
     tRowMergerInit(&merge, piRow, pReader->pSchema);
-    doMergeRowsInBuf(&pBlockScanInfo->iiter, ik.ts, pBlockScanInfo->delSkyline, &merge, pReader);
+    doMergeRowsInBuf(&pBlockScanInfo->iiter, pBlockScanInfo->uid, ik.ts, pBlockScanInfo->delSkyline, &merge, pReader);
 
     tRowMerge(&merge, pRow);
-    doMergeRowsInBuf(&pBlockScanInfo->iter, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
+    doMergeRowsInBuf(&pBlockScanInfo->iter, pBlockScanInfo->uid, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
   } else {
     updateSchema(pRow, pBlockScanInfo->uid, pReader);
 
     tRowMergerInit(&merge, pRow, pReader->pSchema);
-    doMergeRowsInBuf(&pBlockScanInfo->iter, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
+    doMergeRowsInBuf(&pBlockScanInfo->iter, pBlockScanInfo->uid, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
 
     tRowMerge(&merge, piRow);
-    doMergeRowsInBuf(&pBlockScanInfo->iiter, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
+    doMergeRowsInBuf(&pBlockScanInfo->iiter, pBlockScanInfo->uid, k.ts, pBlockScanInfo->delSkyline, &merge, pReader);
   }
 
   tRowMergerGetRow(&merge, pTSRow);
@@ -2592,7 +2617,8 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
     STsdbReader*    pPrevReader = pReader->innerReader[0];
     SDataBlockIter* pBlockIter = &pPrevReader->status.blockIter;
 
-    initFilesetIterator(&pPrevReader->status.fileIter, pPrevReader->pReadSnap->fs.aDFileSet, pPrevReader->order, pPrevReader->idStr);
+    initFilesetIterator(&pPrevReader->status.fileIter, pPrevReader->pReadSnap->fs.aDFileSet, pPrevReader->order,
+                        pPrevReader->idStr);
     resetDataBlockIterator(&pPrevReader->status.blockIter, pPrevReader->order, pReader->status.pTableMap);
 
     // no data in files, let's try buffer in memory
@@ -2647,12 +2673,14 @@ void tsdbReaderClose(STsdbReader* pReader) {
 
   SIOCostSummary* pCost = &pReader->cost;
 
-  tsdbDebug("%p :io-cost summary: head-file:%" PRIu64 ", head-file time:%.2f ms, SMA:%"PRId64" SMA-time:%.2f ms, "
-            "fileBlocks:%"PRId64", fileBlocks-time:%.2f ms, build in-memory-block-time:%.2f ms, STableBlockScanInfo "
+  tsdbDebug("%p :io-cost summary: head-file:%" PRIu64 ", head-file time:%.2f ms, SMA:%" PRId64
+            " SMA-time:%.2f ms, "
+            "fileBlocks:%" PRId64
+            ", fileBlocks-time:%.2f ms, build in-memory-block-time:%.2f ms, STableBlockScanInfo "
             "size:%.2f Kb %s",
             pReader, pCost->headFileLoad, pCost->headFileLoadTime, pCost->smaData, pCost->smaLoadTime,
             pCost->numOfBlocks, pCost->blockLoadTime, pCost->buildmemBlock,
-            numOfTables * sizeof(STableBlockScanInfo) /1000.0, pReader->idStr);
+            numOfTables * sizeof(STableBlockScanInfo) / 1000.0, pReader->idStr);
 
   taosMemoryFree(pReader->idStr);
   taosMemoryFree(pReader->pSchema);
@@ -2747,7 +2775,7 @@ int32_t tsdbRetrieveDatablockSMA(STsdbReader* pReader, SColumnDataAgg*** pBlockS
   int32_t code = 0;
   *allHave = false;
 
-  if(pReader->type == TIMEWINDOW_RANGE_EXTERNAL) {
+  if (pReader->type == TIMEWINDOW_RANGE_EXTERNAL) {
     *pBlockStatis = NULL;
     return TSDB_CODE_SUCCESS;
   }
@@ -2891,8 +2919,8 @@ int32_t tsdbReaderReset(STsdbReader* pReader, SQueryTableDataCond* pCond) {
   } else {
     code = initForFirstBlockInFile(pReader, pBlockIter);
     if (code != TSDB_CODE_SUCCESS) {
-      tsdbError("%p reset reader failed, numOfTables:%d, query range:%" PRId64 " - %" PRId64 " in query %s",
-                pReader, numOfTables, pReader->window.skey, pReader->window.ekey, pReader->idStr);
+      tsdbError("%p reset reader failed, numOfTables:%d, query range:%" PRId64 " - %" PRId64 " in query %s", pReader,
+                numOfTables, pReader->window.skey, pReader->window.ekey, pReader->idStr);
       return code;
     }
   }

source/dnode/vnode/src/tsdb/tsdbUtil.c

@@ -13,6 +13,7 @@
  * along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include "tdataformat.h"
 #include "tsdb.h"
 
 // SMapData =======================================================================
@@ -548,6 +549,103 @@ SColVal *tRowIterNext(SRowIter *pIter) {
 }
 
 // SRowMerger ======================================================
+
+int32_t tRowMergerInit2(SRowMerger *pMerger, STSchema *pResTSchema, TSDBROW *pRow, STSchema *pTSchema) {
+  int32_t   code = 0;
+  TSDBKEY   key = TSDBROW_KEY(pRow);
+  SColVal  *pColVal = &(SColVal){0};
+  STColumn *pTColumn;
+  int32_t   iCol, jCol = 0;
+
+  pMerger->pTSchema = pResTSchema;
+  pMerger->version = key.version;
+
+  pMerger->pArray = taosArrayInit(pResTSchema->numOfCols, sizeof(SColVal));
+  if (pMerger->pArray == NULL) {
+    code = TSDB_CODE_OUT_OF_MEMORY;
+    goto _exit;
+  }
+
+  // ts
+  pTColumn = &pTSchema->columns[jCol++];
+
+  ASSERT(pTColumn->type == TSDB_DATA_TYPE_TIMESTAMP);
+
+  *pColVal = COL_VAL_VALUE(pTColumn->colId, pTColumn->type, (SValue){.ts = key.ts});
+  if (taosArrayPush(pMerger->pArray, pColVal) == NULL) {
+    code = TSDB_CODE_OUT_OF_MEMORY;
+    goto _exit;
+  }
+
+  // other
+  for (iCol = 1; jCol < pTSchema->numOfCols && iCol < pResTSchema->numOfCols; ++iCol) {
+    pTColumn = &pResTSchema->columns[iCol];
+    if (pTSchema->columns[jCol].colId < pTColumn->colId) {
+      ++jCol;
+      --iCol;
+      continue;
+    } else if (pTSchema->columns[jCol].colId > pTColumn->colId) {
+      taosArrayPush(pMerger->pArray, &COL_VAL_NONE(pTColumn->colId, pTColumn->type));
+      continue;
+    }
+
+    tsdbRowGetColVal(pRow, pTSchema, jCol++, pColVal);
+    if (taosArrayPush(pMerger->pArray, pColVal) == NULL) {
+      code = TSDB_CODE_OUT_OF_MEMORY;
+      goto _exit;
+    }
+  }
+
+  for (; iCol < pResTSchema->numOfCols; ++iCol) {
+    pTColumn = &pResTSchema->columns[iCol];
+    taosArrayPush(pMerger->pArray, &COL_VAL_NONE(pTColumn->colId, pTColumn->type));
+  }
+
+_exit:
+  return code;
+}
+
+int32_t tRowMergerAdd(SRowMerger *pMerger, TSDBROW *pRow, STSchema *pTSchema) {
+  int32_t   code = 0;
+  TSDBKEY   key = TSDBROW_KEY(pRow);
+  SColVal  *pColVal = &(SColVal){0};
+  STColumn *pTColumn;
+  int32_t   iCol, jCol = 1;
+
+  ASSERT(((SColVal *)pMerger->pArray->pData)->value.ts == key.ts);
+
+  for (iCol = 1; iCol < pMerger->pTSchema->numOfCols && jCol < pTSchema->numOfCols; ++iCol) {
+    pTColumn = &pMerger->pTSchema->columns[iCol];
+    if (pTSchema->columns[jCol].colId < pTColumn->colId) {
+      ++jCol;
+      --iCol;
+      continue;
+    } else if (pTSchema->columns[jCol].colId > pTColumn->colId) {
+      continue;
+    }
+
+    tsdbRowGetColVal(pRow, pTSchema, jCol++, pColVal);
+
+    if (key.version > pMerger->version) {
+      if (!pColVal->isNone) {
+        taosArraySet(pMerger->pArray, iCol, pColVal);
+      }
+    } else if (key.version < pMerger->version) {
+      SColVal *tColVal = (SColVal *)taosArrayGet(pMerger->pArray, iCol);
+      if (tColVal->isNone && !pColVal->isNone) {
+        taosArraySet(pMerger->pArray, iCol, pColVal);
+      }
+    } else {
+      ASSERT(0);
+    }
+  }
+
+  pMerger->version = key.version;
+
+_exit:
+  return code;
+}
+
 int32_t tRowMergerInit(SRowMerger *pMerger, TSDBROW *pRow, STSchema *pTSchema) {
   int32_t   code = 0;
   TSDBKEY   key = TSDBROW_KEY(pRow);

source/dnode/vnode/src/vnd/vnodeSvr.c

@@ -297,8 +297,8 @@ int32_t vnodeProcessQueryMsg(SVnode *pVnode, SRpcMsg *pMsg) {
 
 int32_t vnodeProcessFetchMsg(SVnode *pVnode, SRpcMsg *pMsg, SQueueInfo *pInfo) {
   vTrace("message in fetch queue is processing");
-  if ((pMsg->msgType == TDMT_SCH_FETCH || pMsg->msgType == TDMT_VND_TABLE_META ||
-       pMsg->msgType == TDMT_VND_TABLE_CFG || pMsg->msgType == TDMT_VND_BATCH_META) &&
+  if ((pMsg->msgType == TDMT_SCH_FETCH || pMsg->msgType == TDMT_VND_TABLE_META || pMsg->msgType == TDMT_VND_TABLE_CFG ||
+       pMsg->msgType == TDMT_VND_BATCH_META) &&
       !vnodeIsLeader(pVnode)) {
     vnodeRedirectRpcMsg(pVnode, pMsg);
     return 0;
@@ -977,6 +977,9 @@ static int32_t vnodeProcessAlterHashRangeReq(SVnode *pVnode, int64_t version, vo
 
 static int32_t vnodeProcessAlterConfigReq(SVnode *pVnode, int64_t version, void *pReq, int32_t len, SRpcMsg *pRsp) {
   SAlterVnodeReq alterReq = {0};
+  bool           walChanged = false;
+  bool           tsdbChanged = false;
+
   if (tDeserializeSAlterVnodeReq(pReq, len, &alterReq) != 0) {
     terrno = TSDB_CODE_INVALID_MSG;
     return TSDB_CODE_INVALID_MSG;
@@ -986,9 +989,54 @@ static int32_t vnodeProcessAlterConfigReq(SVnode *pVnode, int64_t version, void
         alterReq.cacheLastSize);
   if (pVnode->config.cacheLastSize != alterReq.cacheLastSize) {
     pVnode->config.cacheLastSize = alterReq.cacheLastSize;
-    // TODO: save config
     tsdbCacheSetCapacity(pVnode, (size_t)pVnode->config.cacheLastSize * 1024 * 1024);
   }
+
+  if (pVnode->config.cacheLast != alterReq.cacheLast) {
+    pVnode->config.cacheLast = alterReq.cacheLast;
+  }
+
+  if (pVnode->config.walCfg.fsyncPeriod != alterReq.walFsyncPeriod) {
+    pVnode->config.walCfg.fsyncPeriod = alterReq.walFsyncPeriod;
+
+    walChanged = true;
+  }
+
+  if (pVnode->config.walCfg.level != alterReq.walLevel) {
+    pVnode->config.walCfg.level = alterReq.walLevel;
+
+    walChanged = true;
+  }
+
+  if (pVnode->config.tsdbCfg.keep0 != alterReq.daysToKeep0) {
+    pVnode->config.tsdbCfg.keep0 = alterReq.daysToKeep0;
+    if (!VND_IS_RSMA(pVnode)) {
+      tsdbChanged = true;
+    }
+  }
+
+  if (pVnode->config.tsdbCfg.keep1 != alterReq.daysToKeep1) {
+    pVnode->config.tsdbCfg.keep1 = alterReq.daysToKeep1;
+    if (!VND_IS_RSMA(pVnode)) {
+      tsdbChanged = true;
+    }
+  }
+
+  if (pVnode->config.tsdbCfg.keep2 != alterReq.daysToKeep2) {
+    pVnode->config.tsdbCfg.keep2 = alterReq.daysToKeep2;
+    if (!VND_IS_RSMA(pVnode)) {
+      tsdbChanged = true;
+    }
+  }
+
+  if (walChanged) {
+    walAlter(pVnode->pWal, &pVnode->config.walCfg);
+  }
+
+  if (tsdbChanged) {
+    tsdbSetKeepCfg(pVnode->pTsdb, &pVnode->config.tsdbCfg);
+  }
+
   return 0;
 }
 

source/libs/catalog/inc/catalogInt.h

@@ -679,6 +679,8 @@ void    ctgClearSubTaskRes(SCtgSubRes *pRes);
 void    ctgFreeQNode(SCtgQNode *node);
 void    ctgClearHandle(SCatalog* pCtg);
 void    ctgFreeTbCacheImpl(SCtgTbCache *pCache);
+int32_t ctgRemoveTbMeta(SCatalog* pCtg, SName* pTableName);
+int32_t ctgGetTbHashVgroup(SCatalog *pCtg, SRequestConnInfo *pConn, const SName *pTableName, SVgroupInfo *pVgroup);
 
 
 extern SCatalogMgmt gCtgMgmt;

source/libs/catalog/src/catalog.c

@@ -92,7 +92,7 @@ int32_t ctgRefreshTbMeta(SCatalog* pCtg, SRequestConnInfo *pConn, SCtgTbMetaCtx*
   int32_t     code = 0;
 
   if (!CTG_FLAG_IS_SYS_DB(ctx->flag)) {
-    CTG_ERR_RET(catalogGetTableHashVgroup(pCtg, pConn, ctx->pName, &vgroupInfo));
+    CTG_ERR_RET(ctgGetTbHashVgroup(pCtg, pConn, ctx->pName, &vgroupInfo));
   }
 
   STableMetaOutput  moutput = {0};
@@ -337,7 +337,10 @@ int32_t ctgGetTbType(SCatalog* pCtg, SRequestConnInfo *pConn, SName* pTableName,
   }
 
   STableMeta* pMeta = NULL;
-  CTG_ERR_RET(catalogGetTableMeta(pCtg, pConn, pTableName, &pMeta));
+  SCtgTbMetaCtx ctx = {0};
+  ctx.pName = (SName*)pTableName;
+  ctx.flag = CTG_FLAG_UNKNOWN_STB;
+  CTG_ERR_RET(ctgGetTbMeta(pCtg, pConn, &ctx, &pMeta));
 
   *tbType = pMeta->tableType;
   taosMemoryFree(pMeta);
@@ -391,7 +394,7 @@ int32_t ctgGetTbCfg(SCatalog* pCtg, SRequestConnInfo *pConn, SName* pTableName,
     CTG_ERR_RET(ctgGetTableCfgFromMnode(pCtg, pConn, pTableName, pCfg, NULL));
   } else {
     SVgroupInfo vgroupInfo = {0};
-    CTG_ERR_RET(catalogGetTableHashVgroup(pCtg, pConn, pTableName, &vgroupInfo));
+    CTG_ERR_RET(ctgGetTbHashVgroup(pCtg, pConn, pTableName, &vgroupInfo));
     CTG_ERR_RET(ctgGetTableCfgFromVnode(pCtg, pConn, pTableName, &vgroupInfo, pCfg, NULL));
   }
 
@@ -477,6 +480,57 @@ _return:
   CTG_RET(code);
 }
 
+
+int32_t ctgGetTbHashVgroup(SCatalog *pCtg, SRequestConnInfo *pConn, const SName *pTableName, SVgroupInfo *pVgroup) {
+  if (IS_SYS_DBNAME(pTableName->dbname)) {
+    ctgError("no valid vgInfo for db, dbname:%s", pTableName->dbname);
+    CTG_ERR_RET(TSDB_CODE_CTG_INVALID_INPUT);
+  }
+
+  SCtgDBCache* dbCache = NULL;
+  int32_t      code = 0;
+  char         db[TSDB_DB_FNAME_LEN] = {0};
+  tNameGetFullDbName(pTableName, db);
+
+  SDBVgInfo *vgInfo = NULL;
+  CTG_ERR_JRET(ctgGetDBVgInfo(pCtg, pConn, db, &dbCache, &vgInfo));
+
+  CTG_ERR_JRET(ctgGetVgInfoFromHashValue(pCtg, vgInfo ? vgInfo : dbCache->vgCache.vgInfo, pTableName, pVgroup));
+
+_return:
+
+  if (dbCache) {
+    ctgRUnlockVgInfo(dbCache);
+    ctgReleaseDBCache(pCtg, dbCache);
+  }
+
+  if (vgInfo) {
+    taosHashCleanup(vgInfo->vgHash);
+    taosMemoryFreeClear(vgInfo);
+  }
+
+  CTG_RET(code);
+}
+
+int32_t ctgRemoveTbMeta(SCatalog* pCtg, SName* pTableName) {
+  int32_t code = 0;
+
+  if (NULL == pCtg || NULL == pTableName) {
+    CTG_ERR_RET(TSDB_CODE_CTG_INVALID_INPUT);
+  }
+
+  if (NULL == pCtg->dbCache) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  CTG_ERR_JRET(ctgRemoveTbMetaFromCache(pCtg, pTableName, true));
+
+_return:
+
+  CTG_RET(code);
+}
+
+
 int32_t catalogInit(SCatalogCfg* cfg) {
   if (gCtgMgmt.pCluster) {
     qError("catalog already initialized");
@@ -772,21 +826,7 @@ _return:
 int32_t catalogRemoveTableMeta(SCatalog* pCtg, SName* pTableName) {
   CTG_API_ENTER();
 
-  int32_t code = 0;
-
-  if (NULL == pCtg || NULL == pTableName) {
-    CTG_API_LEAVE(TSDB_CODE_CTG_INVALID_INPUT);
-  }
-
-  if (NULL == pCtg->dbCache) {
-    CTG_API_LEAVE(TSDB_CODE_SUCCESS);
-  }
-
-  CTG_ERR_JRET(ctgRemoveTbMetaFromCache(pCtg, pTableName, true));
-
-_return:
-
-  CTG_API_LEAVE(code);
+  CTG_API_LEAVE(ctgRemoveTbMeta(pCtg, pTableName));
 }
 
 int32_t catalogRemoveStbMeta(SCatalog* pCtg, const char* dbFName, uint64_t dbId, const char* stbName, uint64_t suid) {
@@ -878,12 +918,12 @@ int32_t catalogChkTbMetaVersion(SCatalog* pCtg, SRequestConnInfo *pConn, SArray*
         case TSDB_CHILD_TABLE: {
           SName stb = name;
           strcpy(stb.tname, stbName);
-          catalogRemoveTableMeta(pCtg, &stb);
+          ctgRemoveTbMeta(pCtg, &stb);
           break;
         }
         case TSDB_SUPER_TABLE:
         case TSDB_NORMAL_TABLE:
-          catalogRemoveTableMeta(pCtg, &name);
+          ctgRemoveTbMeta(pCtg, &name);
           break;
         default:
           ctgError("ignore table type %d", tbType);
@@ -947,34 +987,7 @@ int32_t catalogGetTableDistVgInfo(SCatalog* pCtg, SRequestConnInfo *pConn, const
 int32_t catalogGetTableHashVgroup(SCatalog *pCtg, SRequestConnInfo *pConn, const SName *pTableName, SVgroupInfo *pVgroup) {
   CTG_API_ENTER();
 
-  if (IS_SYS_DBNAME(pTableName->dbname)) {
-    ctgError("no valid vgInfo for db, dbname:%s", pTableName->dbname);
-    CTG_API_LEAVE(TSDB_CODE_CTG_INVALID_INPUT);
-  }
-
-  SCtgDBCache* dbCache = NULL;
-  int32_t      code = 0;
-  char         db[TSDB_DB_FNAME_LEN] = {0};
-  tNameGetFullDbName(pTableName, db);
-
-  SDBVgInfo *vgInfo = NULL;
-  CTG_ERR_JRET(ctgGetDBVgInfo(pCtg, pConn, db, &dbCache, &vgInfo));
-
-  CTG_ERR_JRET(ctgGetVgInfoFromHashValue(pCtg, vgInfo ? vgInfo : dbCache->vgCache.vgInfo, pTableName, pVgroup));
-
-_return:
-
-  if (dbCache) {
-    ctgRUnlockVgInfo(dbCache);
-    ctgReleaseDBCache(pCtg, dbCache);
-  }
-
-  if (vgInfo) {
-    taosHashCleanup(vgInfo->vgHash);
-    taosMemoryFreeClear(vgInfo);
-  }
-
-  CTG_API_LEAVE(code);
+  CTG_API_LEAVE(ctgGetTbHashVgroup(pCtg, pConn, pTableName, pVgroup));
 }
 
 int32_t catalogGetAllMeta(SCatalog* pCtg, SRequestConnInfo *pConn, const SCatalogReq* pReq, SMetaData* pRsp) {
@@ -1200,7 +1213,7 @@ int32_t catalogRefreshGetTableCfg(SCatalog* pCtg, SRequestConnInfo *pConn, const
   }
 
   int32_t code = 0;
-  CTG_ERR_JRET(catalogRemoveTableMeta(pCtg, (SName*)pTableName));
+  CTG_ERR_JRET(ctgRemoveTbMeta(pCtg, (SName*)pTableName));
 
   CTG_ERR_JRET(ctgGetTbCfg(pCtg, pConn, (SName*)pTableName, pCfg));
 

source/libs/catalog/src/ctgAsync.c

@@ -398,7 +398,7 @@ int32_t ctgHandleForceUpdate(SCatalog* pCtg, int32_t taskNum, SCtgJob *pJob, con
 
   SName* name = taosHashIterate(pTb, NULL);
   while (name) {
-    catalogRemoveTableMeta(pCtg, name);
+    ctgRemoveTbMeta(pCtg, name);
     name = taosHashIterate(pTb, name);
   }
 

source/libs/command/inc/commandInt.h

@@ -63,6 +63,8 @@ extern "C" {
 #define EXPLAIN_EXEC_TIME_FORMAT "Execution Time: %.3f ms"
 
 //append area
+#define EXPLAIN_LIMIT_FORMAT "limit=%" PRId64
+#define EXPLAIN_SLIMIT_FORMAT "slimit=%" PRId64
 #define EXPLAIN_LEFT_PARENTHESIS_FORMAT " ("
 #define EXPLAIN_RIGHT_PARENTHESIS_FORMAT ")"
 #define EXPLAIN_BLANK_FORMAT " "
@@ -81,6 +83,8 @@ extern "C" {
 #define EXPLAIN_STRING_TYPE_FORMAT "%s"
 #define EXPLAIN_INPUT_ORDER_FORMAT "input_order=%s"
 #define EXPLAIN_OUTPUT_ORDER_TYPE_FORMAT "output_order=%s"
+#define EXPLAIN_OFFSET_FORMAT "offset=%d"
+#define EXPLAIN_SOFFSET_FORMAT "soffset=%d"
 
 #define COMMAND_RESET_LOG "resetLog"
 #define COMMAND_SCHEDULE_POLICY "schedulePolicy"
@@ -147,6 +151,21 @@ typedef struct SExplainCtx {
 #define EXPLAIN_SUM_ROW_NEW(...) tlen = snprintf(tbuf + VARSTR_HEADER_SIZE, TSDB_EXPLAIN_RESULT_ROW_SIZE - VARSTR_HEADER_SIZE, __VA_ARGS__)
 #define EXPLAIN_SUM_ROW_END() do { varDataSetLen(tbuf, tlen); tlen += VARSTR_HEADER_SIZE; } while (0)
 
+#define EXPLAIN_ROW_APPEND_LIMIT_IMPL(_pLimit, sl) do {                                            \
+  if (_pLimit) {                                                                                   \
+    EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);                                                      \
+    SLimitNode* pLimit = (SLimitNode*)_pLimit;                                                     \
+    EXPLAIN_ROW_APPEND(((sl) ? EXPLAIN_SLIMIT_FORMAT : EXPLAIN_LIMIT_FORMAT), pLimit->limit);      \
+    if (pLimit->offset) {                                                                          \
+      EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);                                                    \
+      EXPLAIN_ROW_APPEND(((sl) ? EXPLAIN_SOFFSET_FORMAT : EXPLAIN_OFFSET_FORMAT), pLimit->offset);\
+    }                                                                                              \
+  }                                                                                                \
+} while (0)
+
+#define EXPLAIN_ROW_APPEND_LIMIT(_pLimit) EXPLAIN_ROW_APPEND_LIMIT_IMPL(_pLimit, false)
+#define EXPLAIN_ROW_APPEND_SLIMIT(_pLimit) EXPLAIN_ROW_APPEND_LIMIT_IMPL(_pLimit, true)
+
 #ifdef __cplusplus
 }
 #endif

source/libs/command/src/command.c

@@ -517,6 +517,13 @@ static int32_t execAlterLocal(SAlterLocalStmt* pStmt) {
     goto _return;
   }
 
+  bool forbidden = false;
+  taosLocalCfgForbiddenToChange(pStmt->config, &forbidden);
+  if (forbidden) {
+    terrno = TSDB_CODE_OPS_NOT_SUPPORT;
+    return terrno;
+  }
+
   if (cfgSetItem(tsCfg, pStmt->config, pStmt->value, CFG_STYPE_ALTER_CMD)) {
     return terrno;
   }

source/libs/command/src/explain.c

@@ -417,6 +417,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pTagScanNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pTagScanNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pTagScanNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pTagScanNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -523,6 +525,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pTblScanNode->scan.node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pTblScanNode->scan.node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pTblScanNode->scan.node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pTblScanNode->scan.node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -567,6 +571,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pSTblScanNode->scan.node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pSTblScanNode->scan.node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pSTblScanNode->scan.node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pSTblScanNode->scan.node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -601,6 +607,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pPrjNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pPrjNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pPrjNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pPrjNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -635,6 +643,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pJoinNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pJoinNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pJoinNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pJoinNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -681,6 +691,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pAggNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pAggNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pAggNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pAggNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -717,6 +729,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pIndefNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pIndefNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pIndefNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pIndefNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -755,6 +769,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pExchNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pExchNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pExchNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pExchNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -826,6 +842,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pSortNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pSortNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pSortNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pSortNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -864,6 +882,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pIntNode->window.node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pIntNode->window.node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pIntNode->window.node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pIntNode->window.node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
         uint8_t precision = getIntervalPrecision(pIntNode);
@@ -906,6 +926,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pIntNode->window.node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pIntNode->window.node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pIntNode->window.node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pIntNode->window.node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
         uint8_t precision = getIntervalPrecision(pIntNode);
@@ -948,6 +970,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pFillNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pFillNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pFillNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pFillNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
         if (pFillNode->pValues) {
@@ -1005,6 +1029,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pSessNode->window.node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pSessNode->window.node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pSessNode->window.node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pSessNode->window.node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -1046,6 +1072,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pStateNode->window.node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pStateNode->window.node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pStateNode->window.node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pStateNode->window.node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -1084,6 +1112,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pPartNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pPartNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pPartNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pPartNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -1153,6 +1183,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pMergeNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pMergeNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pMergeNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pMergeNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -1202,6 +1234,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pDistScanNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pDistScanNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pDistScanNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pDistScanNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -1241,6 +1275,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pLastRowNode->scan.node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pLastRowNode->scan.node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pLastRowNode->scan.node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pLastRowNode->scan.node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -1310,6 +1346,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pSortNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pSortNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pSortNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pSortNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
 
@@ -1342,6 +1380,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pIntNode->window.node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pIntNode->window.node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pIntNode->window.node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pIntNode->window.node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
         uint8_t precision = getIntervalPrecision(pIntNode);
@@ -1389,6 +1429,8 @@ int32_t qExplainResNodeToRowsImpl(SExplainResNode *pResNode, SExplainCtx *ctx, i
                            nodesGetOutputNumFromSlotList(pInterpNode->node.pOutputDataBlockDesc->pSlots));
         EXPLAIN_ROW_APPEND(EXPLAIN_BLANK_FORMAT);
         EXPLAIN_ROW_APPEND(EXPLAIN_WIDTH_FORMAT, pInterpNode->node.pOutputDataBlockDesc->outputRowSize);
+        EXPLAIN_ROW_APPEND_LIMIT(pInterpNode->node.pLimit);
+        EXPLAIN_ROW_APPEND_SLIMIT(pInterpNode->node.pSlimit);
         EXPLAIN_ROW_END();
         QRY_ERR_RET(qExplainResAppendRow(ctx, tbuf, tlen, level + 1));
         if (pInterpNode->pFillValues) {

source/libs/executor/inc/executorimpl.h

@@ -549,7 +549,8 @@ typedef struct SMergeAlignedIntervalAggOperatorInfo {
   SIntervalAggOperatorInfo *intervalAggOperatorInfo;
 
   bool         hasGroupId;
-  uint64_t     groupId;
+  uint64_t     groupId;  // current groupId
+  int64_t      curTs;    // current ts
   SSDataBlock* prefetchedBlock;
   bool         inputBlocksFinished;
 

source/libs/executor/src/executor.c

@@ -48,7 +48,6 @@ static int32_t doSetStreamBlock(SOperatorInfo* pOperator, void* input, size_t nu
     pOperator->status = OP_NOT_OPENED;
 
     SStreamScanInfo* pInfo = pOperator->info;
-    /*pInfo->assignBlockUid = assignUid;*/
 
     // TODO: if a block was set but not consumed,
     // prevent setting a different type of block

source/libs/executor/src/executorimpl.c

@@ -1445,6 +1445,7 @@ static void doUpdateNumOfRows(SResultRow* pRow, int32_t numOfExprs, const int32_
   }
 }
 
+// todo extract method with copytoSSDataBlock
 int32_t finalizeResultRowIntoResultDataBlock(SDiskbasedBuf* pBuf, SResultRowPosition* resultRowPosition,
                                              SqlFunctionCtx* pCtx, SExprInfo* pExprInfo, int32_t numOfExprs,
                                              const int32_t* rowCellOffset, SSDataBlock* pBlock,

source/libs/executor/src/groupoperator.c

@@ -507,6 +507,7 @@ static void doHashPartition(SOperatorInfo* pOperator, SSDataBlock* pBlock) {
           colDataSetNull_f(bitmap, (*rows));
         } else {
           memcpy(data + (*columnLen), colDataGetData(pColInfoData, j), bytes);
+          ASSERT((data + (*columnLen) + bytes - (char*)pPage) <= getBufPageSize(pInfo->pBuf));
         }
         contentLen = bytes;
       }

source/libs/executor/src/scanoperator.c

@@ -481,10 +481,6 @@ static SSDataBlock* doTableScanImpl(SOperatorInfo* pOperator) {
       pBlock->info.groupId = *groupId;
     }
 
-    if (pTableScanInfo->assignBlockUid) {
-      pBlock->info.groupId = pBlock->info.uid;
-    }
-
     pOperator->resultInfo.totalRows = pTableScanInfo->readRecorder.totalRows;
     pTableScanInfo->readRecorder.elapsedTime += (taosGetTimestampUs() - st) / 1000.0;
 
@@ -1174,12 +1170,6 @@ static int32_t setBlockIntoRes(SStreamScanInfo* pInfo, const SSDataBlock* pBlock
     pInfo->pRes->info.groupId = 0;
   }
 
-  // for generating rollup SMA result, each time is an independent time serie.
-  // TODO temporarily used, when the statement of "partition by tbname" is ready, remove this
-  if (pInfo->assignBlockUid) {
-    pInfo->pRes->info.groupId = pBlock->info.uid;
-  }
-
   // todo extract method
   for (int32_t i = 0; i < taosArrayGetSize(pInfo->pColMatchInfo); ++i) {
     SColMatchInfo* pColMatchInfo = taosArrayGet(pInfo->pColMatchInfo, i);

source/libs/executor/src/timewindowoperator.c

@@ -4589,11 +4589,10 @@ void destroyMergeAlignedIntervalOperatorInfo(void* param, int32_t numOfOutput) {
 static int32_t outputMergeAlignedIntervalResult(SOperatorInfo* pOperatorInfo, uint64_t tableGroupId,
                                                 SSDataBlock* pResultBlock, TSKEY wstartTs) {
   SMergeAlignedIntervalAggOperatorInfo* miaInfo = pOperatorInfo->info;
+
   SIntervalAggOperatorInfo* iaInfo = miaInfo->intervalAggOperatorInfo;
   SExecTaskInfo*            pTaskInfo = pOperatorInfo->pTaskInfo;
-
   SExprSupp*                pSup = &pOperatorInfo->exprSupp;
-  bool       ascScan = (iaInfo->inputOrder == TSDB_ORDER_ASC);
 
   SET_RES_WINDOW_KEY(iaInfo->aggSup.keyBuf, &wstartTs, TSDB_KEYSIZE, tableGroupId);
   SResultRowPosition* p1 = (SResultRowPosition*)taosHashGet(iaInfo->aggSup.pResultRowHashTable, iaInfo->aggSup.keyBuf,
@@ -4603,8 +4602,9 @@ static int32_t outputMergeAlignedIntervalResult(SOperatorInfo* pOperatorInfo, ui
   finalizeResultRowIntoResultDataBlock(iaInfo->aggSup.pResultBuf, p1, pSup->pCtx, pSup->pExprInfo, pSup->numOfExprs,
                                        pSup->rowEntryInfoOffset, pResultBlock, pTaskInfo);
   taosHashRemove(iaInfo->aggSup.pResultRowHashTable, iaInfo->aggSup.keyBuf, GET_RES_WINDOW_KEY_LEN(TSDB_KEYSIZE));
+  ASSERT(taosHashGetSize(iaInfo->aggSup.pResultRowHashTable) == 0);
 
-  return 0;
+  return TSDB_CODE_SUCCESS;
 }
 
 static void doMergeAlignedIntervalAggImpl(SOperatorInfo* pOperatorInfo, SResultRowInfo* pResultRowInfo,
@@ -4619,11 +4619,20 @@ static void doMergeAlignedIntervalAggImpl(SOperatorInfo* pOperatorInfo, SResultR
   int32_t     numOfOutput = pSup->numOfExprs;
   int64_t*    tsCols = extractTsCol(pBlock, iaInfo);
   uint64_t    tableGroupId = pBlock->info.groupId;
-  TSKEY       blockStartTs = getStartTsKey(&pBlock->info.window, tsCols);
+  TSKEY       currTs = getStartTsKey(&pBlock->info.window, tsCols);
   SResultRow* pResult = NULL;
 
-  STimeWindow win;
-  win.skey = blockStartTs;
+  // there is an result exists
+  if (miaInfo->curTs != INT64_MIN) {
+    ASSERT(taosHashGetSize(iaInfo->aggSup.pResultRowHashTable) == 1);
+    if (currTs != miaInfo->curTs) {
+      outputMergeAlignedIntervalResult(pOperatorInfo, tableGroupId, pResultBlock, miaInfo->curTs);
+      miaInfo->curTs = INT64_MIN;
+    }
+  }
+
+  STimeWindow win = {0};
+  win.skey = currTs;
   win.ekey =
       taosTimeAdd(win.skey, iaInfo->interval.interval, iaInfo->interval.intervalUnit, iaInfo->interval.precision) - 1;
 
@@ -4634,41 +4643,48 @@ static void doMergeAlignedIntervalAggImpl(SOperatorInfo* pOperatorInfo, SResultR
     longjmp(pTaskInfo->env, TSDB_CODE_QRY_OUT_OF_MEMORY);
   }
 
-  TSKEY       currTs = blockStartTs;
-  TSKEY       currPos = startPos;
+  miaInfo->curTs = win.skey;
+  int32_t currPos = startPos;
+
   STimeWindow currWin = win;
-  while (1) {
-    ++currPos;
-    if (currPos >= pBlock->info.rows) {
-      break;
-    }
+  while (++currPos < pBlock->info.rows) {
     if (tsCols[currPos] == currTs) {
       continue;
-    } else {
+    }
+
     updateTimeWindowInfo(&iaInfo->twAggSup.timeWindowData, &currWin, true);
     doApplyFunctions(pTaskInfo, pSup->pCtx, &currWin, &iaInfo->twAggSup.timeWindowData, startPos, currPos - startPos,
                      tsCols, pBlock->info.rows, numOfOutput, iaInfo->inputOrder);
 
     outputMergeAlignedIntervalResult(pOperatorInfo, tableGroupId, pResultBlock, currTs);
+    miaInfo->curTs = INT64_MIN;
 
     currTs = tsCols[currPos];
     currWin.skey = currTs;
     currWin.ekey = taosTimeAdd(currWin.skey, iaInfo->interval.interval, iaInfo->interval.intervalUnit,
-                                 iaInfo->interval.precision) -
-                     1;
+                               iaInfo->interval.precision) - 1;
+
     startPos = currPos;
-      ret = setTimeWindowOutputBuf(pResultRowInfo, &currWin, (scanFlag == MAIN_SCAN), &pResult, tableGroupId,
-                                   pSup->pCtx, numOfOutput, pSup->rowEntryInfoOffset, &iaInfo->aggSup, pTaskInfo);
+    ret = setTimeWindowOutputBuf(pResultRowInfo, &currWin, (scanFlag == MAIN_SCAN), &pResult, tableGroupId, pSup->pCtx,
+                                 numOfOutput, pSup->rowEntryInfoOffset, &iaInfo->aggSup, pTaskInfo);
     if (ret != TSDB_CODE_SUCCESS || pResult == NULL) {
       longjmp(pTaskInfo->env, TSDB_CODE_QRY_OUT_OF_MEMORY);
     }
+
+    miaInfo->curTs = currWin.skey;
   }
-  }
+
   updateTimeWindowInfo(&iaInfo->twAggSup.timeWindowData, &currWin, true);
   doApplyFunctions(pTaskInfo, pSup->pCtx, &currWin, &iaInfo->twAggSup.timeWindowData, startPos, currPos - startPos,
                    tsCols, pBlock->info.rows, numOfOutput, iaInfo->inputOrder);
 
+  if (currPos >= pBlock->info.rows) {
+    // we need to see next block if exists
+  } else {
+    ASSERT(0);
     outputMergeAlignedIntervalResult(pOperatorInfo, tableGroupId, pResultBlock, currTs);
+    miaInfo->curTs = INT64_MIN;
+  }
 }
 
 static SSDataBlock* doMergeAlignedIntervalAgg(SOperatorInfo* pOperator) {
@@ -4682,12 +4698,14 @@ static SSDataBlock* doMergeAlignedIntervalAgg(SOperatorInfo* pOperator) {
 
   SExprSupp*   pSup = &pOperator->exprSupp;
   SSDataBlock* pRes = iaInfo->binfo.pRes;
+
   blockDataCleanup(pRes);
   blockDataEnsureCapacity(pRes, pOperator->resultInfo.capacity);
 
   if (!miaInfo->inputBlocksFinished) {
     SOperatorInfo* downstream = pOperator->pDownstream[0];
     int32_t        scanFlag = MAIN_SCAN;
+
     while (1) {
       SSDataBlock* pBlock = NULL;
       if (miaInfo->prefetchedBlock == NULL) {
@@ -4699,6 +4717,14 @@ static SSDataBlock* doMergeAlignedIntervalAgg(SOperatorInfo* pOperator) {
       }
 
       if (pBlock == NULL) {
+        // close last unfinalized time window
+        if (miaInfo->curTs != INT64_MIN) {
+          ASSERT(taosHashGetSize(iaInfo->aggSup.pResultRowHashTable) == 1);
+          outputMergeAlignedIntervalResult(pOperator, miaInfo->groupId, pRes, miaInfo->curTs);
+          miaInfo->curTs = INT64_MIN;
+        }
+
+        doSetOperatorCompleted(pOperator);
         miaInfo->inputBlocksFinished = true;
         break;
       }
@@ -4707,7 +4733,11 @@ static SSDataBlock* doMergeAlignedIntervalAgg(SOperatorInfo* pOperator) {
         miaInfo->hasGroupId = true;
         miaInfo->groupId = pBlock->info.groupId;
       } else if (miaInfo->groupId != pBlock->info.groupId) {
+        // if there are unclosed time window, close it firstly.
+        ASSERT(miaInfo->curTs != INT64_MIN);
+        outputMergeAlignedIntervalResult(pOperator, miaInfo->groupId, pRes, miaInfo->curTs);
         miaInfo->prefetchedBlock = pBlock;
+        miaInfo->curTs = INT64_MIN;
         break;
       }
 
@@ -4722,11 +4752,8 @@ static SSDataBlock* doMergeAlignedIntervalAgg(SOperatorInfo* pOperator) {
 
     pRes->info.groupId = miaInfo->groupId;
   }
-  miaInfo->hasGroupId = false;
 
-  if (miaInfo->inputBlocksFinished) {
-    doSetOperatorCompleted(pOperator);
-  }
+  miaInfo->hasGroupId = false;
 
   size_t rows = pRes->info.rows;
   pOperator->resultInfo.totalRows += rows;
@@ -4752,6 +4779,8 @@ SOperatorInfo* createMergeAlignedIntervalOperatorInfo(SOperatorInfo* downstream,
   SExprSupp*                pSup = &pOperator->exprSupp;
 
   miaInfo->pCondition = pCondition;
+  miaInfo->curTs = INT64_MIN;
+
   iaInfo->win = pTaskInfo->window;
   iaInfo->inputOrder = TSDB_ORDER_ASC;
   iaInfo->interval = *pInterval;

source/libs/function/src/builtins.c

@@ -2236,7 +2236,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
   {
     .name = "interp",
     .type = FUNCTION_TYPE_INTERP,
-    .classification = FUNC_MGT_TIMELINE_FUNC | FUNC_MGT_INTERVAL_INTERPO_FUNC | FUNC_MGT_IMPLICIT_TS_FUNC,
+    .classification = FUNC_MGT_TIMELINE_FUNC | FUNC_MGT_INTERVAL_INTERPO_FUNC | FUNC_MGT_IMPLICIT_TS_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
     .translateFunc = translateFirstLast,
     .getEnvFunc    = getSelectivityFuncEnv,
     .initFunc      = functionSetup,
@@ -2246,7 +2246,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
   {
     .name = "derivative",
     .type = FUNCTION_TYPE_DERIVATIVE,
-    .classification = FUNC_MGT_INDEFINITE_ROWS_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_TIMELINE_FUNC | FUNC_MGT_IMPLICIT_TS_FUNC | FUNC_MGT_CUMULATIVE_FUNC,
+    .classification = FUNC_MGT_INDEFINITE_ROWS_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_TIMELINE_FUNC | FUNC_MGT_IMPLICIT_TS_FUNC | FUNC_MGT_CUMULATIVE_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
     .translateFunc = translateDerivative,
     .getEnvFunc   = getDerivativeFuncEnv,
     .initFunc     = derivativeFuncSetup,
@@ -2280,7 +2280,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
   {
     .name = "_cache_last_row",
     .type = FUNCTION_TYPE_CACHE_LAST_ROW,
-    .classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_MULTI_RES_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_IMPLICIT_TS_FUNC,
+    .classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_MULTI_RES_FUNC | FUNC_MGT_SELECT_FUNC | FUNC_MGT_IMPLICIT_TS_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
     .translateFunc = translateFirstLast,
     .getEnvFunc   = getFirstLastFuncEnv,
     .initFunc     = functionSetup,
@@ -2375,7 +2375,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
   {
     .name = "histogram",
     .type = FUNCTION_TYPE_HISTOGRAM,
-    .classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_TIMELINE_FUNC | FUNC_MGT_MULTI_ROWS_FUNC | FUNC_MGT_FORBID_FILL_FUNC,
+    .classification = FUNC_MGT_AGG_FUNC | FUNC_MGT_TIMELINE_FUNC | FUNC_MGT_MULTI_ROWS_FUNC | FUNC_MGT_FORBID_FILL_FUNC | FUNC_MGT_FORBID_STREAM_FUNC,
     .translateFunc = translateHistogram,
     .getEnvFunc   = getHistogramFuncEnv,
     .initFunc     = histogramFunctionSetup,
@@ -2460,7 +2460,7 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
     .processFunc  = diffFunction,
     .sprocessFunc = diffScalarFunction,
     .finalizeFunc = functionFinalize,
-    .estimateReturnRowsFunc = diffEstReturnRows
+    .estimateReturnRowsFunc = diffEstReturnRows,
   },
   {
     .name = "statecount",
@@ -2521,8 +2521,8 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
   {
     .name = "tail",
     .type = FUNCTION_TYPE_TAIL,
-    .classification = FUNC_MGT_SELECT_FUNC | FUNC_MGT_INDEFINITE_ROWS_FUNC | FUNC_MGT_TIMELINE_FUNC | FUNC_MGT_FORBID_STREAM_FUNC |
-                      FUNC_MGT_IMPLICIT_TS_FUNC,
+    .classification = FUNC_MGT_SELECT_FUNC | FUNC_MGT_INDEFINITE_ROWS_FUNC | FUNC_MGT_TIMELINE_FUNC |
+                      FUNC_MGT_FORBID_STREAM_FUNC | FUNC_MGT_IMPLICIT_TS_FUNC,
     .translateFunc = translateTail,
     .getEnvFunc   = getTailFuncEnv,
     .initFunc     = tailFunctionSetup,

source/libs/index/src/indexFilter.c

@@ -385,6 +385,37 @@ static int32_t sifDoIndex(SIFParam *left, SIFParam *right, int8_t operType, SIFP
                            .reverse = reverse,
                            .filterFunc = filterFunc};
 
+    char   buf[128] = {0};
+    float  f = 0.0;
+    double d = 0.0;
+    if (IS_VAR_DATA_TYPE(left->colValType)) {
+      if (!IS_VAR_DATA_TYPE(right->colValType)) {
+        NUM_TO_STRING(right->colValType, right->condValue, sizeof(buf) - 2, buf + VARSTR_HEADER_SIZE);
+        varDataSetLen(buf, strlen(buf + VARSTR_HEADER_SIZE));
+        param.val = buf;
+      }
+    } else {
+      if (left->colValType == TSDB_DATA_TYPE_FLOAT) {
+        if (right->colValType == TSDB_DATA_TYPE_DOUBLE) {
+          f = GET_DOUBLE_VAL(right->condValue);
+          param.val = &f;
+        } else if (right->colValType == TSDB_DATA_TYPE_BIGINT) {
+          f = *(int64_t *)(right->condValue);
+          param.val = &f;
+        } else {
+          f = *(int32_t *)(right->condValue);
+          param.val = &f;
+        }
+      } else if (left->colValType == TSDB_DATA_TYPE_DOUBLE) {
+        if (right->colValType == TSDB_DATA_TYPE_DOUBLE) {
+          d = GET_DOUBLE_VAL(right->condValue);
+          param.val = &d;
+        } else if (right->colValType == TSDB_DATA_TYPE_BIGINT) {
+          d = *(int64_t *)(right->condValue);
+          param.val = &d;
+        }
+      }
+    }
     ret = metaFilterTableIds(arg->metaEx, &param, output->result);
   }
   return ret;

source/libs/nodes/src/nodesCodeFuncs.c

@@ -2684,6 +2684,7 @@ static int32_t jsonToDataType(const SJson* pJson, void* pObj) {
 
 static const char* jkExprDataType = "DataType";
 static const char* jkExprAliasName = "AliasName";
+static const char* jkExprUserAlias = "UserAlias";
 
 static int32_t exprNodeToJson(const void* pObj, SJson* pJson) {
   const SExprNode* pNode = (const SExprNode*)pObj;
@@ -2692,6 +2693,9 @@ static int32_t exprNodeToJson(const void* pObj, SJson* pJson) {
   if (TSDB_CODE_SUCCESS == code) {
     code = tjsonAddStringToObject(pJson, jkExprAliasName, pNode->aliasName);
   }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddStringToObject(pJson, jkExprUserAlias, pNode->userAlias);
+  }
 
   return code;
 }
@@ -2703,6 +2707,9 @@ static int32_t jsonToExprNode(const SJson* pJson, void* pObj) {
   if (TSDB_CODE_SUCCESS == code) {
     code = tjsonGetStringValue(pJson, jkExprAliasName, pNode->aliasName);
   }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonGetStringValue(pJson, jkExprUserAlias, pNode->userAlias);
+  }
 
   return code;
 }

source/libs/nodes/src/nodesUtilFuncs.c

@@ -1502,7 +1502,7 @@ static EDealRes collectColumns(SNode* pNode, void* pContext) {
   SCollectColumnsCxt* pCxt = (SCollectColumnsCxt*)pContext;
   if (QUERY_NODE_COLUMN == nodeType(pNode)) {
     SColumnNode* pCol = (SColumnNode*)pNode;
-    if (isCollectType(pCxt->collectType, pCol->colType) &&
+    if (isCollectType(pCxt->collectType, pCol->colType) && 0 != strcmp(pCol->colName, "*") &&
         (NULL == pCxt->pTableAlias || 0 == strcmp(pCxt->pTableAlias, pCol->tableAlias))) {
       return doCollect(pCxt, pCol, pNode);
     }
@@ -1816,187 +1816,3 @@ int32_t nodesMergeConds(SNode** pDst, SNodeList** pSrc) {
 
   return TSDB_CODE_SUCCESS;
 }
-
-typedef struct SClassifyConditionCxt {
-  bool hasPrimaryKey;
-  bool hasTagIndexCol;
-  bool hasTagCol;
-  bool hasOtherCol;
-} SClassifyConditionCxt;
-
-static EDealRes classifyConditionImpl(SNode* pNode, void* pContext) {
-  SClassifyConditionCxt* pCxt = (SClassifyConditionCxt*)pContext;
-  if (QUERY_NODE_COLUMN == nodeType(pNode)) {
-    SColumnNode* pCol = (SColumnNode*)pNode;
-    if (PRIMARYKEY_TIMESTAMP_COL_ID == pCol->colId && TSDB_SYSTEM_TABLE != pCol->tableType) {
-      pCxt->hasPrimaryKey = true;
-    } else if (pCol->hasIndex) {
-      pCxt->hasTagIndexCol = true;
-      pCxt->hasTagCol = true;
-    } else if (COLUMN_TYPE_TAG == pCol->colType || COLUMN_TYPE_TBNAME == pCol->colType) {
-      pCxt->hasTagCol = true;
-    } else {
-      pCxt->hasOtherCol = true;
-    }
-  }
-  return DEAL_RES_CONTINUE;
-}
-
-typedef enum EConditionType {
-  COND_TYPE_PRIMARY_KEY = 1,
-  COND_TYPE_TAG_INDEX,
-  COND_TYPE_TAG,
-  COND_TYPE_NORMAL
-} EConditionType;
-
-static EConditionType classifyCondition(SNode* pNode) {
-  SClassifyConditionCxt cxt = {.hasPrimaryKey = false, .hasTagIndexCol = false, .hasOtherCol = false};
-  nodesWalkExpr(pNode, classifyConditionImpl, &cxt);
-  return cxt.hasOtherCol ? COND_TYPE_NORMAL
-                         : (cxt.hasPrimaryKey && cxt.hasTagCol
-                                ? COND_TYPE_NORMAL
-                                : (cxt.hasPrimaryKey ? COND_TYPE_PRIMARY_KEY
-                                                     : (cxt.hasTagIndexCol ? COND_TYPE_TAG_INDEX : COND_TYPE_TAG)));
-}
-
-static int32_t partitionLogicCond(SNode** pCondition, SNode** pPrimaryKeyCond, SNode** pTagIndexCond, SNode** pTagCond,
-                                  SNode** pOtherCond) {
-  SLogicConditionNode* pLogicCond = (SLogicConditionNode*)(*pCondition);
-
-  int32_t code = TSDB_CODE_SUCCESS;
-
-  SNodeList* pPrimaryKeyConds = NULL;
-  SNodeList* pTagIndexConds = NULL;
-  SNodeList* pTagConds = NULL;
-  SNodeList* pOtherConds = NULL;
-  SNode*     pCond = NULL;
-  FOREACH(pCond, pLogicCond->pParameterList) {
-    switch (classifyCondition(pCond)) {
-      case COND_TYPE_PRIMARY_KEY:
-        if (NULL != pPrimaryKeyCond) {
-          code = nodesListMakeAppend(&pPrimaryKeyConds, nodesCloneNode(pCond));
-        }
-        break;
-      case COND_TYPE_TAG_INDEX:
-        if (NULL != pTagIndexCond) {
-          code = nodesListMakeAppend(&pTagIndexConds, nodesCloneNode(pCond));
-        }
-        if (NULL != pTagCond) {
-          code = nodesListMakeAppend(&pTagConds, nodesCloneNode(pCond));
-        }
-        break;
-      case COND_TYPE_TAG:
-        if (NULL != pTagCond) {
-          code = nodesListMakeAppend(&pTagConds, nodesCloneNode(pCond));
-        }
-        break;
-      case COND_TYPE_NORMAL:
-      default:
-        if (NULL != pOtherCond) {
-          code = nodesListMakeAppend(&pOtherConds, nodesCloneNode(pCond));
-        }
-        break;
-    }
-    if (TSDB_CODE_SUCCESS != code) {
-      break;
-    }
-  }
-
-  SNode* pTempPrimaryKeyCond = NULL;
-  SNode* pTempTagIndexCond = NULL;
-  SNode* pTempTagCond = NULL;
-  SNode* pTempOtherCond = NULL;
-  if (TSDB_CODE_SUCCESS == code) {
-    code = nodesMergeConds(&pTempPrimaryKeyCond, &pPrimaryKeyConds);
-  }
-  if (TSDB_CODE_SUCCESS == code) {
-    code = nodesMergeConds(&pTempTagIndexCond, &pTagIndexConds);
-  }
-  if (TSDB_CODE_SUCCESS == code) {
-    code = nodesMergeConds(&pTempTagCond, &pTagConds);
-  }
-  if (TSDB_CODE_SUCCESS == code) {
-    code = nodesMergeConds(&pTempOtherCond, &pOtherConds);
-  }
-
-  if (TSDB_CODE_SUCCESS == code) {
-    if (NULL != pPrimaryKeyCond) {
-      *pPrimaryKeyCond = pTempPrimaryKeyCond;
-    }
-    if (NULL != pTagIndexCond) {
-      *pTagIndexCond = pTempTagIndexCond;
-    }
-    if (NULL != pTagCond) {
-      *pTagCond = pTempTagCond;
-    }
-    if (NULL != pOtherCond) {
-      *pOtherCond = pTempOtherCond;
-    }
-    nodesDestroyNode(*pCondition);
-    *pCondition = NULL;
-  } else {
-    nodesDestroyList(pPrimaryKeyConds);
-    nodesDestroyList(pTagIndexConds);
-    nodesDestroyList(pTagConds);
-    nodesDestroyList(pOtherConds);
-    nodesDestroyNode(pTempPrimaryKeyCond);
-    nodesDestroyNode(pTempTagIndexCond);
-    nodesDestroyNode(pTempTagCond);
-    nodesDestroyNode(pTempOtherCond);
-  }
-
-  return code;
-}
-
-int32_t nodesPartitionCond(SNode** pCondition, SNode** pPrimaryKeyCond, SNode** pTagIndexCond, SNode** pTagCond,
-                           SNode** pOtherCond) {
-  if (QUERY_NODE_LOGIC_CONDITION == nodeType(*pCondition) &&
-      LOGIC_COND_TYPE_AND == ((SLogicConditionNode*)*pCondition)->condType) {
-    return partitionLogicCond(pCondition, pPrimaryKeyCond, pTagIndexCond, pTagCond, pOtherCond);
-  }
-
-  bool needOutput = false;
-  switch (classifyCondition(*pCondition)) {
-    case COND_TYPE_PRIMARY_KEY:
-      if (NULL != pPrimaryKeyCond) {
-        *pPrimaryKeyCond = *pCondition;
-        needOutput = true;
-      }
-      break;
-    case COND_TYPE_TAG_INDEX:
-      if (NULL != pTagIndexCond) {
-        *pTagIndexCond = *pCondition;
-        needOutput = true;
-      }
-      if (NULL != pTagCond) {
-        SNode* pTempCond = *pCondition;
-        if (NULL != pTagIndexCond) {
-          pTempCond = nodesCloneNode(*pCondition);
-          if (NULL == pTempCond) {
-            return TSDB_CODE_OUT_OF_MEMORY;
-          }
-        }
-        *pTagCond = pTempCond;
-        needOutput = true;
-      }
-      break;
-    case COND_TYPE_TAG:
-      if (NULL != pTagCond) {
-        *pTagCond = *pCondition;
-        needOutput = true;
-      }
-      break;
-    case COND_TYPE_NORMAL:
-    default:
-      if (NULL != pOtherCond) {
-        *pOtherCond = *pCondition;
-        needOutput = true;
-      }
-      break;
-  }
-  if (needOutput) {
-    *pCondition = NULL;
-  }
-
-  return TSDB_CODE_SUCCESS;
-}

source/libs/parser/src/parAstCreater.c

@@ -233,18 +233,22 @@ SNode* createRawExprNodeExt(SAstCreateContext* pCxt, const SToken* pStart, const
 SNode* releaseRawExprNode(SAstCreateContext* pCxt, SNode* pNode) {
   CHECK_PARSER_STATUS(pCxt);
   SRawExprNode* pRawExpr = (SRawExprNode*)pNode;
-  SNode*        pExpr = pRawExpr->pNode;
-  if (nodesIsExprNode(pExpr)) {
+  SNode*        pRealizedExpr = pRawExpr->pNode;
+  if (nodesIsExprNode(pRealizedExpr)) {
+    SExprNode* pExpr = (SExprNode*)pRealizedExpr;
     if (QUERY_NODE_COLUMN == nodeType(pExpr)) {
-      strcpy(((SExprNode*)pExpr)->aliasName, ((SColumnNode*)pExpr)->colName);
+      strcpy(pExpr->aliasName, ((SColumnNode*)pExpr)->colName);
+      strcpy(pExpr->userAlias, ((SColumnNode*)pExpr)->colName);
     } else {
-      int32_t len = TMIN(sizeof(((SExprNode*)pExpr)->aliasName) - 1, pRawExpr->n);
-      strncpy(((SExprNode*)pExpr)->aliasName, pRawExpr->p, len);
-      ((SExprNode*)pExpr)->aliasName[len] = '\0';
+      int32_t len = TMIN(sizeof(pExpr->aliasName) - 1, pRawExpr->n);
+      strncpy(pExpr->aliasName, pRawExpr->p, len);
+      pExpr->aliasName[len] = '\0';
+      strncpy(pExpr->userAlias, pRawExpr->p, len);
+      pExpr->userAlias[len] = '\0';
     }
   }
   taosMemoryFreeClear(pNode);
-  return pExpr;
+  return pRealizedExpr;
 }
 
 SToken getTokenFromRawExprNode(SAstCreateContext* pCxt, SNode* pNode) {
@@ -641,11 +645,12 @@ SNode* createInterpTimeRange(SAstCreateContext* pCxt, SNode* pStart, SNode* pEnd
 SNode* setProjectionAlias(SAstCreateContext* pCxt, SNode* pNode, SToken* pAlias) {
   CHECK_PARSER_STATUS(pCxt);
   trimEscape(pAlias);
-  int32_t len = TMIN(sizeof(((SExprNode*)pNode)->aliasName) - 1, pAlias->n);
-  strncpy(((SExprNode*)pNode)->aliasName, pAlias->z, len);
-  ((SExprNode*)pNode)->aliasName[len] = '\0';
-  strncpy(((SExprNode*)pNode)->userAlias, pAlias->z, len);
-  ((SExprNode*)pNode)->userAlias[len] = '\0';
+  SExprNode* pExpr = (SExprNode*)pNode;
+  int32_t    len = TMIN(sizeof(pExpr->aliasName) - 1, pAlias->n);
+  strncpy(pExpr->aliasName, pAlias->z, len);
+  pExpr->aliasName[len] = '\0';
+  strncpy(pExpr->userAlias, pAlias->z, len);
+  pExpr->userAlias[len] = '\0';
   return pNode;
 }
 
@@ -766,13 +771,21 @@ SNode* createSelectStmt(SAstCreateContext* pCxt, bool isDistinct, SNodeList* pPr
   return (SNode*)select;
 }
 
+static void setSubquery(SNode* pStmt) {
+  if (QUERY_NODE_SELECT_STMT == nodeType(pStmt)) {
+    ((SSelectStmt*)pStmt)->isSubquery = true;
+  }
+}
+
 SNode* createSetOperator(SAstCreateContext* pCxt, ESetOperatorType type, SNode* pLeft, SNode* pRight) {
   CHECK_PARSER_STATUS(pCxt);
   SSetOperator* setOp = (SSetOperator*)nodesMakeNode(QUERY_NODE_SET_OPERATOR);
   CHECK_OUT_OF_MEM(setOp);
   setOp->opType = type;
   setOp->pLeft = pLeft;
+  setSubquery(setOp->pLeft);
   setOp->pRight = pRight;
+  setSubquery(setOp->pRight);
   sprintf(setOp->stmtName, "%p", setOp);
   return (SNode*)setOp;
 }

source/libs/parser/src/parTranslater.c

@@ -523,6 +523,9 @@ static void setColumnInfoBySchema(const SRealTableNode* pTable, const SSchema* p
   if ('\0' == pCol->node.aliasName[0]) {
     strcpy(pCol->node.aliasName, pColSchema->name);
   }
+  if ('\0' == pCol->node.userAlias[0]) {
+    strcpy(pCol->node.userAlias, pColSchema->name);
+  }
   pCol->tableId = pTable->pMeta->uid;
   pCol->tableType = pTable->pMeta->tableType;
   pCol->colId = pColSchema->colId;
@@ -549,6 +552,9 @@ static void setColumnInfoByExpr(STempTableNode* pTable, SExprNode* pExpr, SColum
   if ('\0' == pCol->node.aliasName[0]) {
     strcpy(pCol->node.aliasName, pCol->colName);
   }
+  if ('\0' == pCol->node.userAlias[0]) {
+    strcpy(pCol->node.userAlias, pCol->colName);
+  }
   pCol->node.resType = pExpr->resType;
 }
 
@@ -691,7 +697,7 @@ static EDealRes translateColumnUseAlias(STranslateContext* pCxt, SColumnNode** p
   SNode*     pNode;
   FOREACH(pNode, pProjectionList) {
     SExprNode* pExpr = (SExprNode*)pNode;
-    if (0 == strcmp((*pCol)->colName, pExpr->aliasName)) {
+    if (0 == strcmp((*pCol)->colName, pExpr->userAlias)) {
       SColumnRefNode* pColRef = (SColumnRefNode*)nodesMakeNode(QUERY_NODE_COLUMN_REF);
       if (NULL == pColRef) {
         pCxt->errCode = TSDB_CODE_OUT_OF_MEMORY;
@@ -1535,6 +1541,7 @@ static EDealRes rewriteColToSelectValFunc(STranslateContext* pCxt, SNode** pNode
   }
   strcpy(pFunc->functionName, "_select_value");
   strcpy(pFunc->node.aliasName, ((SExprNode*)*pNode)->aliasName);
+  strcpy(pFunc->node.userAlias, ((SExprNode*)*pNode)->userAlias);
   pCxt->errCode = nodesListMakeAppend(&pFunc->pParameterList, *pNode);
   if (TSDB_CODE_SUCCESS == pCxt->errCode) {
     pCxt->errCode = getFuncInfo(pCxt, pFunc);
@@ -2171,12 +2178,51 @@ static int32_t translateFillValues(STranslateContext* pCxt, SSelectStmt* pSelect
   return TSDB_CODE_SUCCESS;
 }
 
+static int32_t rewriteProjectAlias(SNodeList* pProjectionList) {
+  int32_t no = 1;
+  SNode*  pProject = NULL;
+  FOREACH(pProject, pProjectionList) {
+    SExprNode* pExpr = (SExprNode*)pProject;
+    if ('\0' == pExpr->userAlias[0]) {
+      strcpy(pExpr->userAlias, pExpr->aliasName);
+    }
+    sprintf(pExpr->aliasName, "#expr_%d", no++);
+  }
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t checkProjectAlias(STranslateContext* pCxt, SNodeList* pProjectionList) {
+  SHashObj* pUserAliasSet = taosHashInit(LIST_LENGTH(pProjectionList),
+                                         taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY), false, HASH_NO_LOCK);
+  SNode*    pProject = NULL;
+  FOREACH(pProject, pProjectionList) {
+    SExprNode* pExpr = (SExprNode*)pProject;
+    if (NULL != taosHashGet(pUserAliasSet, pExpr->userAlias, strlen(pExpr->userAlias))) {
+      taosHashCleanup(pUserAliasSet);
+      return generateSyntaxErrMsg(&pCxt->msgBuf, TSDB_CODE_PAR_AMBIGUOUS_COLUMN, pExpr->userAlias);
+    }
+    taosHashPut(pUserAliasSet, pExpr->userAlias, strlen(pExpr->userAlias), &pExpr, POINTER_BYTES);
+  }
+  taosHashCleanup(pUserAliasSet);
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t translateProjectionList(STranslateContext* pCxt, SSelectStmt* pSelect) {
+  if (pSelect->isSubquery) {
+    return checkProjectAlias(pCxt, pSelect->pProjectionList);
+  }
+  return rewriteProjectAlias(pSelect->pProjectionList);
+}
+
 static int32_t translateSelectList(STranslateContext* pCxt, SSelectStmt* pSelect) {
   pCxt->currClause = SQL_CLAUSE_SELECT;
   int32_t code = translateExprList(pCxt, pSelect->pProjectionList);
   if (TSDB_CODE_SUCCESS == code) {
     code = translateStar(pCxt, pSelect);
   }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = translateProjectionList(pCxt, pSelect);
+  }
   if (TSDB_CODE_SUCCESS == code) {
     code = checkExprListForGroupBy(pCxt, pSelect, pSelect->pProjectionList);
   }
@@ -2232,7 +2278,7 @@ static int32_t getQueryTimeRange(STranslateContext* pCxt, SNode* pWhere, STimeWi
   }
 
   SNode* pPrimaryKeyCond = NULL;
-  nodesPartitionCond(&pCond, &pPrimaryKeyCond, NULL, NULL, NULL);
+  filterPartitionCond(&pCond, &pPrimaryKeyCond, NULL, NULL, NULL);
 
   int32_t code = TSDB_CODE_SUCCESS;
   if (NULL != pPrimaryKeyCond) {
@@ -2699,6 +2745,7 @@ static SNode* createSetOperProject(const char* pTableAlias, SNode* pNode) {
   strcpy(pCol->tableAlias, pTableAlias);
   strcpy(pCol->colName, ((SExprNode*)pNode)->aliasName);
   strcpy(pCol->node.aliasName, pCol->colName);
+  strcpy(pCol->node.userAlias, ((SExprNode*)pNode)->userAlias);
   return (SNode*)pCol;
 }
 
@@ -2810,7 +2857,7 @@ static int32_t partitionDeleteWhere(STranslateContext* pCxt, SDeleteStmt* pDelet
 
   SNode*  pPrimaryKeyCond = NULL;
   SNode*  pOtherCond = NULL;
-  int32_t code = nodesPartitionCond(&pDelete->pWhere, &pPrimaryKeyCond, NULL, &pDelete->pTagCond, &pOtherCond);
+  int32_t code = filterPartitionCond(&pDelete->pWhere, &pPrimaryKeyCond, NULL, &pDelete->pTagCond, &pOtherCond);
   if (TSDB_CODE_SUCCESS == code && NULL != pOtherCond) {
     code = generateSyntaxErrMsg(&pCxt->msgBuf, TSDB_CODE_PAR_INVALID_DELETE_WHERE);
   }

source/libs/parser/test/parSelectTest.cpp

@@ -70,6 +70,8 @@ TEST_F(ParserSelectTest, condition) {
   run("SELECT c1 FROM t1 WHERE NOT ts in (true, false)");
 
   run("SELECT * FROM t1 WHERE c1 > 10 and c1 is not null");
+
+  run("SELECT * FROM t1 WHERE TBNAME like 'fda%' or TS > '2021-05-05 18:19:01.000'");
 }
 
 TEST_F(ParserSelectTest, pseudoColumn) {

source/libs/planner/src/planOptimizer.c

@@ -436,7 +436,7 @@ static int32_t pushDownCondOptDealScan(SOptimizeContext* pCxt, SScanLogicNode* p
 
   SNode*  pPrimaryKeyCond = NULL;
   SNode*  pOtherCond = NULL;
-  int32_t code = nodesPartitionCond(&pScan->node.pConditions, &pPrimaryKeyCond, &pScan->pTagIndexCond, &pScan->pTagCond,
+  int32_t code = filterPartitionCond(&pScan->node.pConditions, &pPrimaryKeyCond, &pScan->pTagIndexCond, &pScan->pTagCond,
                                     &pOtherCond);
   if (TSDB_CODE_SUCCESS == code && NULL != pScan->pTagCond) {
     code = pushDownCondOptRebuildTbanme(&pScan->pTagCond);

source/libs/planner/src/planSpliter.c

@@ -248,12 +248,25 @@ static bool stbSplNeedSplitWindow(bool streamQuery, SLogicNode* pNode) {
   return false;
 }
 
+static bool stbSplNeedSplitJoin(bool streamQuery, SJoinLogicNode* pJoin) {
+  if (pJoin->isSingleTableJoin) {
+    return false;
+  }
+  SNode* pChild = NULL;
+  FOREACH(pChild, pJoin->node.pChildren) {
+    if (QUERY_NODE_LOGIC_PLAN_SCAN != nodeType(pChild) && QUERY_NODE_LOGIC_PLAN_JOIN != nodeType(pChild)) {
+      return false;
+    }
+  }
+  return true;
+}
+
 static bool stbSplNeedSplit(bool streamQuery, SLogicNode* pNode) {
   switch (nodeType(pNode)) {
     case QUERY_NODE_LOGIC_PLAN_SCAN:
       return stbSplIsMultiTbScan(streamQuery, (SScanLogicNode*)pNode);
     case QUERY_NODE_LOGIC_PLAN_JOIN:
-      return !(((SJoinLogicNode*)pNode)->isSingleTableJoin);
+      return stbSplNeedSplitJoin(streamQuery, (SJoinLogicNode*)pNode);
     case QUERY_NODE_LOGIC_PLAN_PARTITION:
       return stbSplIsMultiTbScanChild(streamQuery, pNode);
     case QUERY_NODE_LOGIC_PLAN_AGG:
@@ -763,6 +776,8 @@ static SNode* stbSplCreateColumnNode(SExprNode* pExpr) {
     return NULL;
   }
   if (QUERY_NODE_COLUMN == nodeType(pExpr)) {
+    strcpy(pCol->dbName, ((SColumnNode*)pExpr)->dbName);
+    strcpy(pCol->tableName, ((SColumnNode*)pExpr)->tableName);
     strcpy(pCol->tableAlias, ((SColumnNode*)pExpr)->tableAlias);
   }
   strcpy(pCol->colName, pExpr->aliasName);

source/libs/planner/test/planOrderByTest.cpp

@@ -39,6 +39,8 @@ TEST_F(PlanOrderByTest, expr) {
   useDb("root", "test");
 
   run("SELECT * FROM t1 ORDER BY c1 + 10, c2");
+
+  run("SELECT c1 FROM st1 ORDER BY ts, _C0");
 }
 
 TEST_F(PlanOrderByTest, nullsOrder) {

source/libs/planner/test/planSubqueryTest.cpp

@@ -73,3 +73,9 @@ TEST_F(PlanSubqeuryTest, outerInterval) {
 
   run("SELECT COUNT(*) FROM (SELECT ts, TOP(c1, 10) FROM st1s1) INTERVAL(5s)");
 }
+
+TEST_F(PlanSubqeuryTest, outerPartition) {
+  useDb("root", "test");
+
+  run("SELECT c1, COUNT(*) FROM (SELECT ts, c1 FROM st1) PARTITION BY c1");
+}

source/libs/scalar/src/filter.c

@@ -22,6 +22,7 @@
 #include "tcompare.h"
 #include "tdatablock.h"
 #include "ttime.h"
+#include "functionMgt.h"
 
 OptrStr gOptrStr[] = {
   {0,                                      "invalid"},
@@ -3877,4 +3878,195 @@ bool filterExecute(SFilterInfo *info, SSDataBlock *pSrc, int8_t** p, SColumnData
 }
 
 
+typedef struct SClassifyConditionCxt {
+  bool hasPrimaryKey;
+  bool hasTagIndexCol;
+  bool hasTagCol;
+  bool hasOtherCol;
+} SClassifyConditionCxt;
+
+static EDealRes classifyConditionImpl(SNode* pNode, void* pContext) {
+  SClassifyConditionCxt* pCxt = (SClassifyConditionCxt*)pContext;
+  if (QUERY_NODE_COLUMN == nodeType(pNode)) {
+    SColumnNode* pCol = (SColumnNode*)pNode;
+    if (PRIMARYKEY_TIMESTAMP_COL_ID == pCol->colId && TSDB_SYSTEM_TABLE != pCol->tableType) {
+      pCxt->hasPrimaryKey = true;
+    } else if (pCol->hasIndex) {
+      pCxt->hasTagIndexCol = true;
+      pCxt->hasTagCol = true;
+    } else if (COLUMN_TYPE_TAG == pCol->colType || COLUMN_TYPE_TBNAME == pCol->colType) {
+      pCxt->hasTagCol = true;
+    } else {
+      pCxt->hasOtherCol = true;
+    }
+  } else if (QUERY_NODE_FUNCTION == nodeType(pNode)) {
+    SFunctionNode* pFunc = (SFunctionNode*)pNode;
+    if (fmIsPseudoColumnFunc(pFunc->funcId)) {
+      if (FUNCTION_TYPE_TBNAME==pFunc->funcType) {
+        pCxt->hasTagCol = true;
+      } else {
+        pCxt->hasOtherCol = true;
+      }
+    }
+  }
+  return DEAL_RES_CONTINUE;
+}
+
+typedef enum EConditionType {
+  COND_TYPE_PRIMARY_KEY = 1,
+  COND_TYPE_TAG_INDEX,
+  COND_TYPE_TAG,
+  COND_TYPE_NORMAL
+} EConditionType;
+
+static EConditionType classifyCondition(SNode* pNode) {
+  SClassifyConditionCxt cxt = {.hasPrimaryKey = false, .hasTagIndexCol = false, .hasOtherCol = false};
+  nodesWalkExpr(pNode, classifyConditionImpl, &cxt);
+  return cxt.hasOtherCol ? COND_TYPE_NORMAL
+                         : (cxt.hasPrimaryKey && cxt.hasTagCol
+                                ? COND_TYPE_NORMAL
+                                : (cxt.hasPrimaryKey ? COND_TYPE_PRIMARY_KEY
+                                                     : (cxt.hasTagIndexCol ? COND_TYPE_TAG_INDEX : COND_TYPE_TAG)));
+}
+
+static int32_t partitionLogicCond(SNode** pCondition, SNode** pPrimaryKeyCond, SNode** pTagIndexCond, SNode** pTagCond,
+                                  SNode** pOtherCond) {
+  SLogicConditionNode* pLogicCond = (SLogicConditionNode*)(*pCondition);
+
+  int32_t code = TSDB_CODE_SUCCESS;
+
+  SNodeList* pPrimaryKeyConds = NULL;
+  SNodeList* pTagIndexConds = NULL;
+  SNodeList* pTagConds = NULL;
+  SNodeList* pOtherConds = NULL;
+  SNode*     pCond = NULL;
+  FOREACH(pCond, pLogicCond->pParameterList) {
+    switch (classifyCondition(pCond)) {
+      case COND_TYPE_PRIMARY_KEY:
+        if (NULL != pPrimaryKeyCond) {
+          code = nodesListMakeAppend(&pPrimaryKeyConds, nodesCloneNode(pCond));
+        }
+        break;
+      case COND_TYPE_TAG_INDEX:
+        if (NULL != pTagIndexCond) {
+          code = nodesListMakeAppend(&pTagIndexConds, nodesCloneNode(pCond));
+        }
+        if (NULL != pTagCond) {
+          code = nodesListMakeAppend(&pTagConds, nodesCloneNode(pCond));
+        }
+        break;
+      case COND_TYPE_TAG:
+        if (NULL != pTagCond) {
+          code = nodesListMakeAppend(&pTagConds, nodesCloneNode(pCond));
+        }
+        break;
+      case COND_TYPE_NORMAL:
+      default:
+        if (NULL != pOtherCond) {
+          code = nodesListMakeAppend(&pOtherConds, nodesCloneNode(pCond));
+        }
+        break;
+    }
+    if (TSDB_CODE_SUCCESS != code) {
+      break;
+    }
+  }
 
+  SNode* pTempPrimaryKeyCond = NULL;
+  SNode* pTempTagIndexCond = NULL;
+  SNode* pTempTagCond = NULL;
+  SNode* pTempOtherCond = NULL;
+  if (TSDB_CODE_SUCCESS == code) {
+    code = nodesMergeConds(&pTempPrimaryKeyCond, &pPrimaryKeyConds);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = nodesMergeConds(&pTempTagIndexCond, &pTagIndexConds);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = nodesMergeConds(&pTempTagCond, &pTagConds);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = nodesMergeConds(&pTempOtherCond, &pOtherConds);
+  }
+
+  if (TSDB_CODE_SUCCESS == code) {
+    if (NULL != pPrimaryKeyCond) {
+      *pPrimaryKeyCond = pTempPrimaryKeyCond;
+    }
+    if (NULL != pTagIndexCond) {
+      *pTagIndexCond = pTempTagIndexCond;
+    }
+    if (NULL != pTagCond) {
+      *pTagCond = pTempTagCond;
+    }
+    if (NULL != pOtherCond) {
+      *pOtherCond = pTempOtherCond;
+    }
+    nodesDestroyNode(*pCondition);
+    *pCondition = NULL;
+  } else {
+    nodesDestroyList(pPrimaryKeyConds);
+    nodesDestroyList(pTagIndexConds);
+    nodesDestroyList(pTagConds);
+    nodesDestroyList(pOtherConds);
+    nodesDestroyNode(pTempPrimaryKeyCond);
+    nodesDestroyNode(pTempTagIndexCond);
+    nodesDestroyNode(pTempTagCond);
+    nodesDestroyNode(pTempOtherCond);
+  }
+
+  return code;
+}
+
+int32_t filterPartitionCond(SNode** pCondition, SNode** pPrimaryKeyCond, SNode** pTagIndexCond, SNode** pTagCond,
+                           SNode** pOtherCond) {
+  if (QUERY_NODE_LOGIC_CONDITION == nodeType(*pCondition) &&
+      LOGIC_COND_TYPE_AND == ((SLogicConditionNode*)*pCondition)->condType) {
+    return partitionLogicCond(pCondition, pPrimaryKeyCond, pTagIndexCond, pTagCond, pOtherCond);
+  }
+
+  bool needOutput = false;
+  switch (classifyCondition(*pCondition)) {
+    case COND_TYPE_PRIMARY_KEY:
+      if (NULL != pPrimaryKeyCond) {
+        *pPrimaryKeyCond = *pCondition;
+        needOutput = true;
+      }
+      break;
+    case COND_TYPE_TAG_INDEX:
+      if (NULL != pTagIndexCond) {
+        *pTagIndexCond = *pCondition;
+        needOutput = true;
+      }
+      if (NULL != pTagCond) {
+        SNode* pTempCond = *pCondition;
+        if (NULL != pTagIndexCond) {
+          pTempCond = nodesCloneNode(*pCondition);
+          if (NULL == pTempCond) {
+            return TSDB_CODE_OUT_OF_MEMORY;
+          }
+        }
+        *pTagCond = pTempCond;
+        needOutput = true;
+      }
+      break;
+    case COND_TYPE_TAG:
+      if (NULL != pTagCond) {
+        *pTagCond = *pCondition;
+        needOutput = true;
+      }
+      break;
+    case COND_TYPE_NORMAL:
+    default:
+      if (NULL != pOtherCond) {
+        *pOtherCond = *pCondition;
+        needOutput = true;
+      }
+      break;
+  }
+  if (needOutput) {
+    *pCondition = NULL;
+  }
+
+  return TSDB_CODE_SUCCESS;
+}

source/libs/scheduler/src/schJob.c

@@ -278,7 +278,7 @@ int32_t schValidateAndBuildJob(SQueryPlan *pDag, SSchJob *pJob) {
   }
 
   SHashObj *planToTask = taosHashInit(
-      SCHEDULE_DEFAULT_MAX_TASK_NUM,
+      pDag->numOfSubplans,
       taosGetDefaultHashFunction(POINTER_BYTES == sizeof(int64_t) ? TSDB_DATA_TYPE_BIGINT : TSDB_DATA_TYPE_INT), false,
       HASH_NO_LOCK);
   if (NULL == planToTask) {

source/libs/stream/src/streamDispatch.c

@@ -462,7 +462,8 @@ int32_t streamDispatch(SStreamTask* pTask) {
   if (streamDispatchAllBlocks(pTask, pBlock) < 0) {
     ASSERT(0);
     code = -1;
-    // TODO set status fail
+    streamQueueProcessFail(pTask->outputQueue);
+    atomic_store_8(&pTask->outputStatus, TASK_OUTPUT_STATUS__NORMAL);
     goto FREE;
   }
   /*atomic_store_8(&pTask->outputStatus, TASK_OUTPUT_STATUS__NORMAL);*/

source/libs/sync/inc/syncInt.h

@@ -191,8 +191,10 @@ int32_t syncNodeStopElectTimer(SSyncNode* pSyncNode);
 int32_t syncNodeRestartElectTimer(SSyncNode* pSyncNode, int32_t ms);
 int32_t syncNodeResetElectTimer(SSyncNode* pSyncNode);
 int32_t syncNodeStartHeartbeatTimer(SSyncNode* pSyncNode);
+int32_t syncNodeStartNowHeartbeatTimer(SSyncNode* pSyncNode);
 int32_t syncNodeStopHeartbeatTimer(SSyncNode* pSyncNode);
 int32_t syncNodeRestartHeartbeatTimer(SSyncNode* pSyncNode);
+int32_t syncNodeRestartNowHeartbeatTimer(SSyncNode* pSyncNode);
 
 // utils --------------
 int32_t syncNodeSendMsgById(const SRaftId* destRaftId, SSyncNode* pSyncNode, SRpcMsg* pMsg);

source/libs/sync/src/syncAppendEntriesReply.c

@@ -104,6 +104,16 @@ int32_t syncNodeOnAppendEntriesReplyCb(SSyncNode* ths, SyncAppendEntriesReply* p
 // only start once
 static void syncNodeStartSnapshotOnce(SSyncNode* ths, SyncIndex beginIndex, SyncIndex endIndex, SyncTerm lastApplyTerm,
                                       SyncAppendEntriesReply* pMsg) {
+  if (beginIndex > endIndex) {
+    do {
+      char logBuf[128];
+      snprintf(logBuf, sizeof(logBuf), "snapshot param error, start:%ld, end:%ld", beginIndex, endIndex);
+      syncNodeErrorLog(ths, logBuf);
+    } while (0);
+
+    return;
+  }
+
   // get sender
   SSyncSnapshotSender* pSender = syncNodeGetSnapshotSender(ths, &(pMsg->srcId));
   ASSERT(pSender != NULL);

source/libs/sync/src/syncMain.c

@@ -1312,6 +1312,17 @@ int32_t syncNodeStartHeartbeatTimer(SSyncNode* pSyncNode) {
   return ret;
 }
 
+int32_t syncNodeStartNowHeartbeatTimer(SSyncNode* pSyncNode) {
+  int32_t ret = 0;
+  if (syncEnvIsStart()) {
+    taosTmrReset(pSyncNode->FpHeartbeatTimerCB, 1, pSyncNode, gSyncEnv->pTimerManager, &pSyncNode->pHeartbeatTimer);
+    atomic_store_64(&pSyncNode->heartbeatTimerLogicClock, pSyncNode->heartbeatTimerLogicClockUser);
+  } else {
+    sError("vgId:%d, start heartbeat timer error, sync env is stop", pSyncNode->vgId);
+  }
+  return ret;
+}
+
 int32_t syncNodeStopHeartbeatTimer(SSyncNode* pSyncNode) {
   int32_t ret = 0;
   atomic_add_fetch_64(&pSyncNode->heartbeatTimerLogicClockUser, 1);
@@ -1326,6 +1337,12 @@ int32_t syncNodeRestartHeartbeatTimer(SSyncNode* pSyncNode) {
   return 0;
 }
 
+int32_t syncNodeRestartNowHeartbeatTimer(SSyncNode* pSyncNode) {
+  syncNodeStopHeartbeatTimer(pSyncNode);
+  syncNodeStartNowHeartbeatTimer(pSyncNode);
+  return 0;
+}
+
 // utils --------------
 int32_t syncNodeSendMsgById(const SRaftId* destRaftId, SSyncNode* pSyncNode, SRpcMsg* pMsg) {
   SEpSet epSet;

source/libs/sync/src/syncReplication.c

@@ -140,6 +140,7 @@ int32_t syncNodeAppendEntriesPeersSnapshot2(SSyncNode* pSyncNode) {
              ", match-index:%d, raftid:%" PRId64,
              pSyncNode->vgId, nextIndex, newNextIndex, SYNC_INDEX_INVALID, pDestId->addr);
 
+      syncNodeRestartNowHeartbeatTimer(pSyncNode);
       return -1;
     }
 

source/libs/wal/src/walRead.c

@@ -441,9 +441,12 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
     return -1;
   }
 
+  taosThreadMutexLock(&pReader->mutex);
+
   if (pReader->curInvalid || pReader->curVersion != ver) {
     if (walReadSeekVer(pReader, ver) < 0) {
       wError("vgId:%d, unexpected wal log, index:%" PRId64 ", since %s", pReader->pWal->cfg.vgId, ver, terrstr());
+      taosThreadMutexUnlock(&pReader->mutex);
       return -1;
     }
     seeked = true;
@@ -464,6 +467,7 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
         terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
       }
       ASSERT(0);
+      taosThreadMutexUnlock(&pReader->mutex);
       return -1;
     }
   }
@@ -473,6 +477,7 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
     wError("vgId:%d, unexpected wal log, index:%" PRId64 ", since head checksum not passed", pReader->pWal->cfg.vgId,
            ver);
     terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
+    taosThreadMutexUnlock(&pReader->mutex);
     return -1;
   }
 
@@ -480,6 +485,7 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
     void *ptr = taosMemoryRealloc(pReader->pHead, sizeof(SWalCkHead) + pReader->pHead->head.bodyLen);
     if (ptr == NULL) {
       terrno = TSDB_CODE_WAL_OUT_OF_MEMORY;
+      taosThreadMutexUnlock(&pReader->mutex);
       return -1;
     }
     pReader->pHead = ptr;
@@ -494,6 +500,7 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
       terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
       ASSERT(0);
     }
+    taosThreadMutexUnlock(&pReader->mutex);
     return -1;
   }
 
@@ -503,6 +510,7 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
     pReader->curInvalid = 1;
     terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
     ASSERT(0);
+    taosThreadMutexUnlock(&pReader->mutex);
     return -1;
   }
 
@@ -516,9 +524,12 @@ int32_t walReadVer(SWalReader *pReader, int64_t ver) {
     pReader->curInvalid = 1;
     terrno = TSDB_CODE_WAL_FILE_CORRUPTED;
     ASSERT(0);
+    taosThreadMutexUnlock(&pReader->mutex);
     return -1;
   }
   pReader->curVersion++;
 
+  taosThreadMutexUnlock(&pReader->mutex);
+
   return 0;
 }

source/libs/wal/src/walWrite.c

@@ -408,7 +408,7 @@ static FORCE_INLINE int32_t walWriteImpl(SWal *pWal, int64_t index, tmsg_t msgTy
   pWal->writeHead.head.version = index;
   pWal->writeHead.head.bodyLen = bodyLen;
   pWal->writeHead.head.msgType = msgType;
-  pWal->writeHead.head.ingestTs = taosGetTimestampMs();
+  pWal->writeHead.head.ingestTs = 0;
 
   // sync info for sync module
   pWal->writeHead.head.syncMeta = syncMeta;

source/os/src/osString.c

@@ -134,21 +134,95 @@ int32_t taosUcs4ToMbs(TdUcs4 *ucs4, int32_t ucs4_max_len, char *mbs) {
 #endif
 }
 
+typedef struct {
+  iconv_t conv;
+  int8_t  inUse;
+} SConv;
+
+SConv *gConv = NULL;
+int32_t convUsed = 0;
+int32_t gConvMaxNum = 0;
+
+void taosConvInit(void) {
+  gConvMaxNum = 512;
+  gConv = taosMemoryCalloc(gConvMaxNum, sizeof(SConv));
+  for (int32_t i = 0; i < gConvMaxNum; ++i) {
+    gConv[i].conv = iconv_open(DEFAULT_UNICODE_ENCODEC, tsCharset);
+    if ((iconv_t)-1 == gConv[i].conv || (iconv_t)0 == gConv[i].conv) {
+      ASSERT(0);
+    }
+  }
+}
+
+void taosConvDestroy() {
+  for (int32_t i = 0; i < gConvMaxNum; ++i) {
+    iconv_close(gConv[i].conv);
+  }
+  taosMemoryFreeClear(gConv);
+  gConvMaxNum = -1;
+}
+
+iconv_t taosAcquireConv(int32_t *idx) {
+  if (gConvMaxNum <= 0) {
+    *idx = -1;
+    return iconv_open(DEFAULT_UNICODE_ENCODEC, tsCharset);
+  }
+  
+  while (true) {
+    int32_t used = atomic_add_fetch_32(&convUsed, 1);
+    if (used > gConvMaxNum) {
+      used = atomic_sub_fetch_32(&convUsed, 1);
+      sched_yield();
+      continue;
+    }
+    
+    break;
+  }
+
+  int32_t startId = taosGetSelfPthreadId() % gConvMaxNum;
+  while (true) {
+    if (gConv[startId].inUse) {
+      startId = (startId + 1) % gConvMaxNum;
+      continue;
+    }
+
+    int8_t old = atomic_val_compare_exchange_8(&gConv[startId].inUse, 0, 1);
+    if (0 == old) {
+      break;
+    }
+  }
+
+  *idx = startId;
+  return gConv[startId].conv;
+}
+
+void taosReleaseConv(int32_t idx, iconv_t conv) {
+  if (idx < 0) {
+    iconv_close(conv);
+    return;
+  }
+
+  atomic_store_8(&gConv[idx].inUse, 0);
+  atomic_sub_fetch_32(&convUsed, 1);
+}
+
 bool taosMbsToUcs4(const char *mbs, size_t mbsLength, TdUcs4 *ucs4, int32_t ucs4_max_len, int32_t *len) {
 #ifdef DISALLOW_NCHAR_WITHOUT_ICONV
   printf("Nchar cannot be read and written without iconv, please install iconv library and recompile TDengine.\n");
   return -1;
 #else
   memset(ucs4, 0, ucs4_max_len);
-  iconv_t cd = iconv_open(DEFAULT_UNICODE_ENCODEC, tsCharset);
+
+  int32_t idx = -1;
+  iconv_t conv = taosAcquireConv(&idx);
   size_t  ucs4_input_len = mbsLength;
   size_t  outLeft = ucs4_max_len;
-  if (iconv(cd, (char **)&mbs, &ucs4_input_len, (char **)&ucs4, &outLeft) == -1) {
-    iconv_close(cd);
+  if (iconv(conv, (char **)&mbs, &ucs4_input_len, (char **)&ucs4, &outLeft) == -1) {
+    taosReleaseConv(idx, conv);
     return false;
   }
 
-  iconv_close(cd);
+  taosReleaseConv(idx, conv);
   if (len != NULL) {
     *len = (int32_t)(ucs4_max_len - outLeft);
     if (*len < 0) {

tests/script/tsim/parser/condition_query.sim

@@ -37,7 +37,7 @@ sql select * from stb1 where c8 < 'nuLl';
 sql select * from stb1 where c9 > 'nuLl';
 sql_error select * from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b;
 sql_error select a.ts,a.c1,a.c8 from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50 or b.c1 < 60;
-sql_error select a.ts,a.c1,a.c8 from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and ((a.c1 > 50 and a.c1 < 60) or (b.c2 > 60));
+sql select a.ts,a.c1,a.c8 from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and ((a.c1 > 50 and a.c1 < 60) or (b.c2 > 60));
 sql select * from stb1 where 'c2' is null;
 sql select * from stb1 where 'c2' is not null;
 
@@ -1475,66 +1475,66 @@ if $data20 != @21-05-05 18:19:25.000@ then
   return -1
 endi
 
-#sql select a.* from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50 order by ts;
-#sql select a.* from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50;
-#if $rows != 4 then
-#  return -1
-#endi
-#if $data00 != @21-05-05 18:19:20.000@ then
-#  return -1
-#endi
-#if $data10 != @21-05-05 18:19:21.000@ then
-#  return -1
-#endi
-#if $data20 != @21-05-05 18:19:24.000@ then
-#  return -1
-#endi
-#if $data30 != @21-05-05 18:19:25.000@ then
-#  return -1
-#endi
+sql select a.* from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50 order by ts;
+sql select a.* from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50;
+if $rows != 4 then
+  return -1
+endi
+if $data00 != @21-05-05 18:19:20.000@ then
+  return -1
+endi
+if $data10 != @21-05-05 18:19:21.000@ then
+  return -1
+endi
+if $data20 != @21-05-05 18:19:24.000@ then
+  return -1
+endi
+if $data30 != @21-05-05 18:19:25.000@ then
+  return -1
+endi
 
-#sql select a.ts,a.c1,a.c8,a.c9 from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50 and b.c1 < 60 order by ts;
-#sql select a.ts,a.c1,a.c8 from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50 and b.c1 < 60 order by ts;
-#if $rows != 2 then
-#  return -1
-#endi
-#if $data00 != @21-05-05 18:19:20.000@ then
-#  return -1
-#endi
-#if $data10 != @21-05-05 18:19:21.000@ then
-#  return -1
-#endi
-#
-#sql select a.ts,b.ts,a.c1,b.u1,b.u2 from (select * from stb1) a, (select * from stb2) b where a.ts=b.ts and (a.c1 < 10 or a.c1 > 30) and (b.u1 < 5 or b.u1 > 5) order by ts;
-#if $rows != 4 then
-#  return -1
-#endi
-#if $data00 != @21-05-05 18:19:00.000@ then
-#  return -1
-#endi
-#if $data10 != @21-05-05 18:19:02.000@ then
-#  return -1
-#endi
-#if $data20 != @21-05-05 18:19:12.000@ then
-#  return -1
-#endi
-#if $data30 != @21-05-05 18:19:14.000@ then
-#  return -1
-#endi
-#
-#sql select a.ts,b.ts,a.c1,b.u1,b.u2 from (select * from stb1) a, (select * from stb2) b where a.ts=b.ts and a.c1 < 30 and b.u1 > 1 and a.c1 > 10 and b.u1 < 8 and b.u1<>5 order by ts;
-#if $rows != 3 then
-#  return -1
-#endi
-#if $data00 != @21-05-05 18:19:04.000@ then
-#  return -1
-#endi
-#if $data10 != @21-05-05 18:19:06.000@ then
-#  return -1
-#endi
-#if $data20 != @21-05-05 18:19:10.000@ then
-#  return -1
-#endi
+sql select a.ts,a.c1,a.c8,a.c9 from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50 and b.c1 < 60 order by ts;
+sql select a.ts,a.c1,a.c8 from (select * from stb1 where c7=true) a, (select * from stb1 where c1 > 30) b where a.ts=b.ts and a.c1 > 50 and b.c1 < 60 order by ts;
+if $rows != 2 then
+  return -1
+endi
+if $data00 != @21-05-05 18:19:20.000@ then
+  return -1
+endi
+if $data10 != @21-05-05 18:19:21.000@ then
+  return -1
+endi
+
+sql select a.ts,b.ts,a.c1,b.u1,b.u2 from (select * from stb1) a, (select * from stb2) b where a.ts=b.ts and (a.c1 < 10 or a.c1 > 30) and (b.u1 < 5 or b.u1 > 5) order by ts;
+if $rows != 4 then
+  return -1
+endi
+if $data00 != @21-05-05 18:19:00.000@ then
+  return -1
+endi
+if $data10 != @21-05-05 18:19:02.000@ then
+  return -1
+endi
+if $data20 != @21-05-05 18:19:12.000@ then
+  return -1
+endi
+if $data30 != @21-05-05 18:19:14.000@ then
+  return -1
+endi
+
+sql select a.ts,b.ts,a.c1,b.u1,b.u2 from (select * from stb1) a, (select * from stb2) b where a.ts=b.ts and a.c1 < 30 and b.u1 > 1 and a.c1 > 10 and b.u1 < 8 and b.u1<>5 order by ts;
+if $rows != 3 then
+  return -1
+endi
+if $data00 != @21-05-05 18:19:04.000@ then
+  return -1
+endi
+if $data10 != @21-05-05 18:19:06.000@ then
+  return -1
+endi
+if $data20 != @21-05-05 18:19:10.000@ then
+  return -1
+endi
 
 sql select * from stb1 where c1 is null and c1 is not null;
 if $rows != 0 then
@@ -2574,7 +2574,7 @@ sql select * from stb1 where c1 > 0 or t1 > 0 and c1 > 1
 sql select * from stb1 where t1 > 0 or t1 > 0 and c1 > 1
 sql select * from stb1 where (c1 > 0 and t1 > 0 ) or (t1 > 1 and c1 > 3)
 sql select * from stb1 where (c1 > 0 and t1 > 0 ) or t1 > 1
-sql_error select a.ts,b.ts,a.c1,b.u1,b.u2 from (select * from stb1) a, (select * from stb2) b where a.ts=b.ts and a.t1=b.t1;
+sql select a.ts,b.ts,a.c1,b.u1,b.u2 from (select * from stb1) a, (select * from stb2) b where a.ts=b.ts and a.t1=b.t1;
 
 sql select * from stb1 where c1 < 63 and t1 > 5
 if $rows != 2 then
@@ -2631,7 +2631,7 @@ sql_error select tb1.ts,tb1.c1,tb2_1.u1 from tb1, tb2_1 where tb1.ts=tb2_1.ts or
 
 
 print "ts&tbname test"
-sql_error select count(*) from stb1 where ts > 0 or tbname like 'tb%';
+sql select count(*) from stb1 where ts > 0 or tbname like 'tb%';
 
 print "ts&tag test"
 sql select count(*) from stb1 where ts > 0 or t1 > 0;
@@ -2717,9 +2717,9 @@ print "tbname&tag&join test"
 
 
 print "column&ts&tbname&tag test"
-sql_error select * from stb1 where (tbname like 'tb%' or ts > '2021-05-05 18:19:01.000') and (t1 > 5 or t1 < 4) and c1 > 0;
+sql select * from stb1 where (tbname like 'tb%' or ts > '2021-05-05 18:19:01.000') and (t1 > 5 or t1 < 4) and c1 > 0;
 sql select * from stb1 where (ts > '2021-05-05 18:19:01.000') and (ts > '2021-05-05 18:19:02.000' or t1 > 3) and (t1 > 5 or t1 < 4) and c1 > 0;
-sql_error select ts,c1,c7 from stb1 where ts > '2021-05-05 18:19:03.000' or ts > '2021-05-05 18:19:20.000' and col > 0 and t1 > 0;
+sql select ts,c1,c7 from stb1 where ts > '2021-05-05 18:19:03.000' or ts > '2021-05-05 18:19:20.000' and c1 > 0 and t1 > 0;
 
 
 print "column&ts&tbname&join test"

tests/script/tsim/parser/first_last_query.sim

@@ -284,7 +284,7 @@ sql connect
 sql use first_db0;
 
 sql select last(*), tbname from m1 group by tbname order by tbname;
-
+print $data00 $data01 $data02 $data10 $data11 $data12
 if $rows != 2 then
   return -1
 endi

tests/script/tsim/query/explain.sim

@@ -45,6 +45,7 @@ print ======== step3
 sql explain verbose true select * from st1 where -2;
 sql explain verbose true select ts from tb1 where f1 > 0;
 sql explain verbose true select * from st1 where f1 > 0 and ts > '2020-10-31 00:00:00' and ts < '2021-10-31 00:00:00';
+sql explain verbose true select count(*) from st1 partition by tbname slimit 1 soffset 2 limit 2 offset 1;
 sql explain verbose true select * from information_schema.user_stables where db_name='db2';
 
 print ======== step4 

tests/system-test/6-cluster/vnode/4dnode1mnode_basic_replica1_insertdatas_querys.py

+# author : wenzhouwww
+from ssl import ALERT_DESCRIPTION_CERTIFICATE_UNOBTAINABLE
+import taos
+import sys
+import time
+import os 
+
+from util.log import *
+from util.sql import *
+from util.cases import *
+from util.dnodes import TDDnodes
+from util.dnodes import TDDnode
+from util.cluster import *
+
+import time
+import socket
+import subprocess ,threading
+sys.path.append(os.path.dirname(__file__))
+
+class TDTestCase:
+    def init(self,conn ,logSql):
+        tdLog.debug(f"start to excute {__file__}")
+        tdSql.init(conn.cursor())
+        self.host = socket.gethostname()
+        self.mnode_list = {}
+        self.dnode_list = {}
+        self.ts = 1483200000000
+        self.db_name ='testdb'
+        self.replica = 1 
+        self.vgroups = 1
+        self.tb_nums = 10 
+        self.row_nums = 1000
+        self.query_times = 500
+
+    def getBuildPath(self):
+        selfPath = os.path.dirname(os.path.realpath(__file__))
+        if ("community" in selfPath):
+            projPath = selfPath[:selfPath.find("community")]
+        else:
+            projPath = selfPath[:selfPath.find("tests")]
+
+        for root, dirs, files in os.walk(projPath):
+            if ("taosd" in files):
+                rootRealPath = os.path.dirname(os.path.realpath(root))
+                if ("packaging" not in rootRealPath):
+                    buildPath = root[:len(root) - len("/build/bin")]
+                    break
+        return buildPath
+
+    def check_setup_cluster_status(self):
+        tdSql.query("show mnodes")
+        for mnode in tdSql.queryResult:
+            name = mnode[1]
+            info = mnode
+            self.mnode_list[name] = info
+
+        tdSql.query("show dnodes")
+        for dnode in tdSql.queryResult:
+            name = dnode[1]
+            info = dnode
+            self.dnode_list[name] = info
+
+        count = 0
+        is_leader = False
+        mnode_name = ''
+        for k,v in self.mnode_list.items():
+            count +=1
+            # only for 1 mnode
+            mnode_name = k
+
+            if v[2] =='leader':
+                is_leader=True
+
+        if count==1 and is_leader:
+            tdLog.notice("===== depoly cluster success with 1 mnode as leader =====")
+        else:
+            tdLog.exit("===== depoly cluster fail with 1 mnode as leader =====")
+
+        for k ,v in self.dnode_list.items():
+            if k == mnode_name:
+                if v[3]==0:
+                    tdLog.notice("===== depoly cluster mnode only success at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+                else:
+                    tdLog.exit("===== depoly cluster mnode only fail at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+            else:
+                continue
+
+    def create_db_check_vgroups(self):
+
+        tdSql.execute("drop database if exists test")
+        tdSql.execute("create database if not exists test replica 1 duration 300")
+        tdSql.execute("use test")
+        tdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        tdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(5):
+            tdSql.execute("create table sub_tb_{} using stb1 tags({})".format(i,i))
+        tdSql.query("show stables")
+        tdSql.checkRows(1)
+        tdSql.query("show tables")
+        tdSql.checkRows(6)
+
+        tdSql.query("show test.vgroups;")
+        vgroups_infos = {}  # key is id: value is info list
+        for vgroup_info in tdSql.queryResult:
+            vgroup_id = vgroup_info[0]
+            tmp_list = []
+            for role in vgroup_info[3:-4]:
+                if role in ['leader','follower']:
+                    tmp_list.append(role)
+            vgroups_infos[vgroup_id]=tmp_list
+
+        for k , v in vgroups_infos.items():
+            if len(v) ==1 and v[0]=="leader":
+                tdLog.notice(" === create database replica only 1 role leader  check success of vgroup_id {} ======".format(k))
+            else:
+                tdLog.exit(" === create database replica only 1 role leader  check fail of vgroup_id {} ======".format(k))
+
+    def create_db_replica_3_insertdatas(self, dbname, replica_num ,vgroup_nums ,tb_nums , row_nums ):
+        newTdSql=tdCom.newTdSql()
+        drop_db_sql = "drop database if exists {}".format(dbname)
+        create_db_sql = "create database {} replica {} vgroups {}".format(dbname,replica_num,vgroup_nums)
+
+        tdLog.notice(" ==== create database {} and insert rows begin =====".format(dbname))
+        newTdSql.execute(drop_db_sql)
+        newTdSql.execute(create_db_sql)
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        newTdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(tb_nums):
+            sub_tbname = "sub_tb_{}".format(i)
+            newTdSql.execute("create table {} using stb1 tags({})".format(sub_tbname,i))
+            # insert datas about new database
+
+            for row_num in range(row_nums):
+                ts = self.ts + 1000*row_num
+                newTdSql.execute(f"insert into {sub_tbname} values ({ts}, {row_num} ,{row_num}, 10 ,1 ,{row_num} ,{row_num},true,'bin_{row_num}','nchar_{row_num}',now) ")
+
+        tdLog.notice(" ==== create database {} and insert rows execute end =====".format(dbname))
+
+    
+
+    def check_insert_status(self, newTdSql ,dbname, tb_nums , row_nums):
+        # newTdSql=tdCom.newTdSql()
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.query("select count(*) from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkData(0 , 0 , tb_nums*row_nums)
+        newTdSql.query("select distinct tbname from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkRows(tb_nums)
+
+    def loop_query_constantly(self, times ,  db_name, tb_nums ,row_nums):
+        newTdSql=tdCom.newTdSql()
+        for loop_time in range(times):
+            tdLog.debug(" === query is going ,this is {}_th query === ".format(loop_time))
+            self.check_insert_status( newTdSql , db_name, tb_nums , row_nums)
+
+    def run(self): 
+        self.check_setup_cluster_status()
+        self.create_db_check_vgroups()
+        
+        # start writing constantly 
+        writing = threading.Thread(target = self.create_db_replica_3_insertdatas, args=(self.db_name , self.replica , self.vgroups , self.tb_nums , self.row_nums))
+        writing.start()
+        tdSql.query(" show {}.stables ".format(self.db_name))
+        while not tdSql.queryResult:
+            print(tdSql.queryResult)
+            time.sleep(0.1)
+            tdSql.query(" show {}.stables ".format(self.db_name))
+
+        reading = threading.Thread(target = self.loop_query_constantly, args=(self.query_times,self.db_name , self.tb_nums , self.row_nums))
+        reading.start()
+        
+        writing.join()
+        reading.join()
+
+    def stop(self):
+        tdSql.close()
+        tdLog.success(f"{__file__} successfully executed")
+
+tdCases.addLinux(__file__, TDTestCase())
+tdCases.addWindows(__file__, TDTestCase())
\ No newline at end of file

tests/system-test/6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_force_stop_all_dnodes.py

+# author : wenzhouwww
+from ssl import ALERT_DESCRIPTION_CERTIFICATE_UNOBTAINABLE
+import taos
+import sys
+import time
+import os 
+
+from util.log import *
+from util.sql import *
+from util.cases import *
+from util.dnodes import TDDnodes
+from util.dnodes import TDDnode
+from util.cluster import *
+
+import datetime
+import inspect
+import time
+import socket
+import subprocess
+import threading
+sys.path.append(os.path.dirname(__file__))
+
+class TDTestCase:
+    def init(self,conn ,logSql):
+        tdLog.debug(f"start to excute {__file__}")
+        tdSql.init(conn.cursor())
+        self.host = socket.gethostname()
+        self.mnode_list = {}
+        self.dnode_list = {}
+        self.ts = 1483200000000
+        self.ts_step =1000
+        self.db_name ='testdb'
+        self.replica = 3 
+        self.vgroups = 1
+        self.tb_nums = 10 
+        self.row_nums = 100
+        self.stop_dnode_id = None
+        self.loop_restart_times = 5
+        self.current_thread = None
+        self.max_restart_time = 10
+        self.try_check_times = 10
+
+    def getBuildPath(self):
+        selfPath = os.path.dirname(os.path.realpath(__file__))
+        if ("community" in selfPath):
+            projPath = selfPath[:selfPath.find("community")]
+        else:
+            projPath = selfPath[:selfPath.find("tests")]
+
+        for root, dirs, files in os.walk(projPath):
+            if ("taosd" in files):
+                rootRealPath = os.path.dirname(os.path.realpath(root))
+                if ("packaging" not in rootRealPath):
+                    buildPath = root[:len(root) - len("/build/bin")]
+                    break
+        return buildPath
+
+    def check_setup_cluster_status(self):
+
+        tdSql.query("show mnodes")
+        for mnode in tdSql.queryResult:
+            name = mnode[1]
+            info = mnode
+            self.mnode_list[name] = info
+
+        tdSql.query("show dnodes")
+        for dnode in tdSql.queryResult:
+            name = dnode[1]
+            info = dnode
+            self.dnode_list[name] = info
+
+        count = 0
+        is_leader = False
+        mnode_name = ''
+        for k,v in self.mnode_list.items():
+            count +=1
+            # only for 1 mnode
+            mnode_name = k
+
+            if v[2] =='leader':
+                is_leader=True
+
+        if count==1 and is_leader:
+            tdLog.notice("===== depoly cluster success with 1 mnode as leader =====")
+        else:
+            tdLog.exit("===== depoly cluster fail with 1 mnode as leader =====")
+
+        for k ,v in self.dnode_list.items():
+            if k == mnode_name:
+                if v[3]==0:
+                    tdLog.notice("===== depoly cluster mnode only success at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+                else:
+                    tdLog.exit("===== depoly cluster mnode only fail at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+            else:
+                continue
+
+    def create_database(self, dbname, replica_num ,vgroup_nums ):
+        drop_db_sql = "drop database if exists {}".format(dbname)
+        create_db_sql = "create database {} replica {} vgroups {}".format(dbname,replica_num,vgroup_nums)
+
+        tdLog.notice(" ==== create database {} and insert rows begin =====".format(dbname))
+        tdSql.execute(drop_db_sql)
+        tdSql.execute(create_db_sql)
+        tdSql.execute("use {}".format(dbname))
+       
+    def create_stable_insert_datas(self,dbname ,stablename , tb_nums , row_nums):
+        tdSql.execute("use {}".format(dbname))
+        tdSql.execute(
+        '''create table {}
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''.format(stablename)
+        )
+    
+        for i in range(tb_nums):
+            sub_tbname = "sub_{}_{}".format(stablename,i)
+            tdSql.execute("create table {} using {} tags({})".format(sub_tbname, stablename ,i))
+            # insert datas about new database
+
+            for row_num in range(row_nums):
+                ts = self.ts + self.ts_step*row_num
+                tdSql.execute(f"insert into {sub_tbname} values ({ts}, {row_num} ,{row_num}, 10 ,1 ,{row_num} ,{row_num},true,'bin_{row_num}','nchar_{row_num}',now) ")
+
+        tdLog.notice(" ==== stable {} insert rows execute end =====".format(stablename))
+    
+    def append_rows_of_exists_tables(self,dbname ,stablename , tbname , append_nums ):
+        
+        tdSql.execute("use {}".format(dbname))
+     
+        for row_num in range(append_nums):
+            tdSql.execute(f"insert into {tbname} values (now, {row_num} ,{row_num}, 10 ,1 ,{row_num} ,{row_num},true,'bin_{row_num}','nchar_{row_num}',now) ")
+            # print(f"insert into {tbname} values (now, {row_num} ,{row_num}, 10 ,1 ,{row_num} ,{row_num},true,'bin_{row_num}','nchar_{row_num}',now) ")
+        tdLog.notice(" ==== append new rows of table {} belongs to  stable {} execute end =====".format(tbname,stablename))
+        os.system("taos -s 'select count(*) from {}.{}';".format(dbname,stablename))
+
+    def check_insert_rows(self, dbname, stablename , tb_nums , row_nums, append_rows):
+    
+        tdSql.execute("use {}".format(dbname))
+        
+        tdSql.query("select count(*) from {}.{}".format(dbname,stablename))
+
+        while not tdSql.queryResult:
+            time.sleep(0.1)
+            tdSql.query("select count(*) from {}.{}".format(dbname,stablename))
+
+        status_OK = self.mycheckData("select count(*) from {}.{}".format(dbname,stablename) ,0 , 0 , tb_nums*row_nums+append_rows)
+        
+        count = 0 
+        while not status_OK :
+            if count > self.try_check_times:
+                os.system("taos -s ' show {}.vgroups; '".format(dbname))
+                tdLog.exit(" ==== check insert rows failed  after {}  try check {} times  of database {}".format(count , self.try_check_times ,dbname))
+                break
+            time.sleep(0.1)
+            tdSql.query("select count(*) from {}.{}".format(dbname,stablename))
+            while not tdSql.queryResult:
+                time.sleep(0.1)
+                tdSql.query("select count(*) from {}.{}".format(dbname,stablename))
+            status_OK = self.mycheckData("select count(*) from {}.{}".format(dbname,stablename) ,0 , 0 , tb_nums*row_nums+append_rows)
+            tdLog.debug(" ==== check insert rows first failed , this is {}_th retry check rows of database {}".format(count , dbname))
+            count += 1
+        
+
+        tdSql.query("select distinct tbname from {}.{}".format(dbname,stablename))
+        while not tdSql.queryResult:
+            time.sleep(0.1)
+            tdSql.query("select distinct tbname from {}.{}".format(dbname,stablename))
+        status_OK = self.mycheckRows("select distinct tbname from {}.{}".format(dbname,stablename) ,tb_nums)
+        count = 0 
+        while not status_OK :
+            if count > self.try_check_times:
+                os.system("taos -s ' show {}.vgroups;'".format(dbname))
+                tdLog.exit(" ==== check insert rows failed  after {}  try check {} times  of database {}".format(count , self.try_check_times ,dbname))
+                break
+            time.sleep(0.1)
+            tdSql.query("select distinct tbname from {}.{}".format(dbname,stablename))
+            while not tdSql.queryResult:
+                time.sleep(0.1)
+                tdSql.query("select distinct tbname from {}.{}".format(dbname,stablename))
+            status_OK = self.mycheckRows("select distinct tbname from {}.{}".format(dbname,stablename) ,tb_nums)
+            tdLog.debug(" ==== check insert tbnames first failed , this is {}_th retry check tbnames of database {}".format(count , dbname))
+            count += 1
+
+    def _get_stop_dnode_id(self,dbname ,dnode_role):
+        tdSql.query("show {}.vgroups".format(dbname))
+        vgroup_infos = tdSql.queryResult
+        status = False
+        for vgroup_info in vgroup_infos:
+            if "error" not in vgroup_info:
+                status = True
+            else:
+                status = False
+        while status!=True :
+            time.sleep(0.1)
+            tdSql.query("show {}.vgroups".format(dbname))
+            vgroup_infos = tdSql.queryResult
+            for vgroup_info in vgroup_infos:
+                if "error" not in vgroup_info:
+                    status = True
+                else:
+                    status = False
+            # print(status)
+        for vgroup_info in vgroup_infos:
+            leader_infos = vgroup_info[3:-4] 
+            # print(vgroup_info)
+            for ind ,role in enumerate(leader_infos):
+                if role == dnode_role:
+                    # print(ind,leader_infos)
+                    self.stop_dnode_id = leader_infos[ind-1]
+                    break
+
+        return self.stop_dnode_id
+
+    def wait_stop_dnode_OK(self ,newTdSql ):
+    
+        def _get_status():
+            # newTdSql=tdCom.newTdSql()
+
+            status =  ""
+            newTdSql.query("show dnodes")
+            dnode_infos = newTdSql.queryResult
+            for dnode_info in dnode_infos:
+                id = dnode_info[0]
+                dnode_status = dnode_info[4]
+                if id == self.stop_dnode_id:
+                    status = dnode_status
+                    break
+            return status 
+
+        status = _get_status()
+        while status !="offline":
+            time.sleep(0.1)
+            status = _get_status()
+            # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
+        tdLog.notice("==== stop_dnode has stopped , id is {} ====".format(self.stop_dnode_id))
+
+    def wait_start_dnode_OK(self , newTdSql):
+    
+        def _get_status():
+            # newTdSql=tdCom.newTdSql()
+            status =  ""
+            newTdSql.query("show dnodes")
+            dnode_infos = newTdSql.queryResult
+            for dnode_info in dnode_infos:
+                id = dnode_info[0]
+                dnode_status = dnode_info[4]
+                if id == self.stop_dnode_id:
+                    status = dnode_status
+                    break
+            return status 
+
+        status = _get_status()
+        while status !="ready":
+            time.sleep(0.1)
+            status = _get_status()
+            # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
+        tdLog.notice("==== stop_dnode has restart , id is {} ====".format(self.stop_dnode_id))
+
+    def _parse_datetime(self,timestr):
+        try:
+            return datetime.datetime.strptime(timestr, '%Y-%m-%d %H:%M:%S.%f')
+        except ValueError:
+            pass
+        try:
+            return datetime.datetime.strptime(timestr, '%Y-%m-%d %H:%M:%S')
+        except ValueError:
+            pass
+
+    def mycheckRowCol(self, sql, row, col):
+        caller = inspect.getframeinfo(inspect.stack()[2][0])
+        if row < 0:
+            args = (caller.filename, caller.lineno, sql, row)
+            tdLog.exit("%s(%d) failed: sql:%s, row:%d is smaller than zero" % args)
+        if col < 0:
+            args = (caller.filename, caller.lineno, sql, row)
+            tdLog.exit("%s(%d) failed: sql:%s, col:%d is smaller than zero" % args)
+        if row > tdSql.queryRows:
+            args = (caller.filename, caller.lineno, sql, row, tdSql.queryRows)
+            tdLog.exit("%s(%d) failed: sql:%s, row:%d is larger than queryRows:%d" % args)
+        if col > tdSql.queryCols:
+            args = (caller.filename, caller.lineno, sql, col, tdSql.queryCols)
+            tdLog.exit("%s(%d) failed: sql:%s, col:%d is larger than queryCols:%d" % args)
+
+    def mycheckData(self, sql ,row, col, data):
+        check_status = True
+        self.mycheckRowCol(sql ,row, col)
+        if tdSql.queryResult[row][col] != data:
+            if tdSql.cursor.istype(col, "TIMESTAMP"):
+                # suppose user want to check nanosecond timestamp if a longer data passed
+                if (len(data) >= 28):
+                    if pd.to_datetime(tdSql.queryResult[row][col]) == pd.to_datetime(data):
+                        tdLog.info("sql:%s, row:%d col:%d data:%d == expect:%s" %
+                            (sql, row, col, tdSql.queryResult[row][col], data))
+                else:
+                    if tdSql.queryResult[row][col] == self._parse_datetime(data):
+                        tdLog.info("sql:%s, row:%d col:%d data:%s == expect:%s" %
+                            (sql, row, col, tdSql.queryResult[row][col], data))
+                return
+
+            if str(tdSql.queryResult[row][col]) == str(data):
+                tdLog.info("sql:%s, row:%d col:%d data:%s == expect:%s" %
+                            (sql, row, col, tdSql.queryResult[row][col], data))
+                return
+            elif isinstance(data, float) and abs(tdSql.queryResult[row][col] - data) <= 0.000001:
+                tdLog.info("sql:%s, row:%d col:%d data:%f == expect:%f" %
+                            (sql, row, col, tdSql.queryResult[row][col], data))
+                return
+            else:
+                caller = inspect.getframeinfo(inspect.stack()[1][0])
+                args = (caller.filename, caller.lineno, sql, row, col, tdSql.queryResult[row][col], data)
+                tdLog.info("%s(%d) failed: sql:%s row:%d col:%d data:%s != expect:%s" % args)
+
+                check_status = False
+
+        if data is None:
+            tdLog.info("sql:%s, row:%d col:%d data:%s == expect:%s" %
+                       (sql, row, col, tdSql.queryResult[row][col], data))
+        elif isinstance(data, str):
+            tdLog.info("sql:%s, row:%d col:%d data:%s == expect:%s" %
+                       (sql, row, col, tdSql.queryResult[row][col], data))
+        # elif isinstance(data, datetime.date):
+        #     tdLog.info("sql:%s, row:%d col:%d data:%s == expect:%s" %
+        #                (sql, row, col, tdSql.queryResult[row][col], data))
+        elif isinstance(data, float):
+            tdLog.info("sql:%s, row:%d col:%d data:%s == expect:%s" %
+                       (sql, row, col, tdSql.queryResult[row][col], data))
+        else:
+            tdLog.info("sql:%s, row:%d col:%d data:%s == expect:%d" %
+                       (sql, row, col, tdSql.queryResult[row][col], data))
+
+        return check_status
+
+    def mycheckRows(self, sql, expectRows):
+        check_status = True
+        if len(tdSql.queryResult) == expectRows:
+            tdLog.info("sql:%s, queryRows:%d == expect:%d" % (sql, len(tdSql.queryResult), expectRows))
+            return True
+        else:
+            caller = inspect.getframeinfo(inspect.stack()[1][0])
+            args = (caller.filename, caller.lineno, sql, len(tdSql.queryResult), expectRows)
+            tdLog.info("%s(%d) failed: sql:%s, queryRows:%d != expect:%d" % args)
+            check_status = False
+        return check_status
+
+    def force_stop_dnode(self, dnode_id ):
+        
+        port = None
+        for k ,v  in self.dnode_list.items():
+            if v[0] == dnode_id:
+                port = k.split(":")[-1]
+            else:
+                continue
+
+        if port:
+            tdLog.notice(" ==== dnode {} will be force stop by kill -9 ====".format(dnode_id))
+            psCmd = '''netstat -anp|grep -w LISTEN|grep -w %s |grep -o "LISTEN.*"|awk '{print $2}'|cut -d/ -f1|head -n1''' %(port)
+            processID = subprocess.check_output(
+                psCmd, shell=True).decode("utf-8")
+            ps_kill_taosd = ''' kill -9 {} '''.format(processID)
+            # print(ps_kill_taosd)
+            os.system(ps_kill_taosd)
+        else :
+            tdLog.exit(" ==== port of dnode {} not found ====".format(dnode_id))
+
+    def stop_All(self):
+
+        tdDnodes = cluster.dnodes
+        # newTdSql=tdCom.newTdSql()
+        # ==== stop all dnode =====
+        for k ,v in self.dnode_list.items():
+            dnode_id = v[0]
+            # tdDnodes[dnode_id-1].stoptaosd()
+            self.force_stop_dnode(dnode_id)
+            # self.wait_stop_dnode_OK(newTdSql)
+
+    def start_All(self):
+        tdDnodes = cluster.dnodes
+        # newTdSql=tdCom.newTdSql()
+        for k ,v in self.dnode_list.items():
+            dnode_id = v[0]
+            start = time.time()
+            tdDnodes[dnode_id-1].starttaosd()
+            # self.wait_start_dnode_OK(newTdSql)
+            end = time.time()
+            time_cost = int(end -start)
+            if time_cost > self.max_restart_time:
+                tdLog.exit(" ==== restart dnode {} cost too much time , please check ====".format(self.stop_dnode_id))
+
+
+        
+    def stop_All_dnodes_check_datas(self):
+
+        # stop follower and insert datas , update tables and create new stables
+        self.create_database(dbname = self.db_name ,replica_num= self.replica  , vgroup_nums= 1)
+        self.create_stable_insert_datas(dbname = self.db_name , stablename = "stb1" , tb_nums= self.tb_nums ,row_nums= self.row_nums )
+        self.check_insert_rows(self.db_name ,'stb1' ,tb_nums = self.tb_nums , row_nums= self.row_nums ,append_rows=0)
+        os.system(" taos -s ' select count(*) from {}.{}  ;'".format(self.db_name ,'stb1' ))
+
+        for i in range(self.loop_restart_times):
+            # begin to stop All taosd 
+            self.stop_All()
+            # begin to start All taosd
+            self.start_All()
+
+            tdLog.debug(" ==== cluster has restart , this is {}_th restart cluster  ==== ".format(i))
+
+            self.check_insert_rows(self.db_name ,'stb1' ,tb_nums = self.tb_nums , row_nums= self.row_nums ,append_rows=0)
+            os.system(" taos -s ' select count(*) from {}.{}  ;'".format(self.db_name ,'stb1' ))
+
+        self.append_rows_of_exists_tables(self.db_name ,"stb1" , "sub_stb1_0" , 100 )
+        os.system(" taos -s ' select count(*) from {}.{}  ;'".format(self.db_name ,'stb1' ))
+            
+
+    def run(self): 
+
+        # basic insert and check of cluster
+        self.check_setup_cluster_status()
+        self.stop_All_dnodes_check_datas()
+        
+        
+
+
+    def stop(self):
+        tdSql.close()
+        tdLog.success(f"{__file__} successfully executed")
+
+tdCases.addLinux(__file__, TDTestCase())
+tdCases.addWindows(__file__, TDTestCase())
\ No newline at end of file

tests/system-test/6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys.py

+# author : wenzhouwww
+from ssl import ALERT_DESCRIPTION_CERTIFICATE_UNOBTAINABLE
+import taos
+import sys
+import time
+import os 
+
+from util.log import *
+from util.sql import *
+from util.cases import *
+from util.dnodes import TDDnodes
+from util.dnodes import TDDnode
+from util.cluster import *
+
+import time
+import socket
+import subprocess ,threading
+sys.path.append(os.path.dirname(__file__))
+
+class TDTestCase:
+    def init(self,conn ,logSql):
+        tdLog.debug(f"start to excute {__file__}")
+        tdSql.init(conn.cursor())
+        self.host = socket.gethostname()
+        self.mnode_list = {}
+        self.dnode_list = {}
+        self.ts = 1483200000000
+        self.db_name ='testdb'
+        self.replica = 3 
+        self.vgroups = 1
+        self.tb_nums = 10 
+        self.row_nums = 1000
+        self.query_times = 1000
+
+    def getBuildPath(self):
+        selfPath = os.path.dirname(os.path.realpath(__file__))
+        if ("community" in selfPath):
+            projPath = selfPath[:selfPath.find("community")]
+        else:
+            projPath = selfPath[:selfPath.find("tests")]
+
+        for root, dirs, files in os.walk(projPath):
+            if ("taosd" in files):
+                rootRealPath = os.path.dirname(os.path.realpath(root))
+                if ("packaging" not in rootRealPath):
+                    buildPath = root[:len(root) - len("/build/bin")]
+                    break
+        return buildPath
+
+    def check_setup_cluster_status(self):
+        tdSql.query("show mnodes")
+        for mnode in tdSql.queryResult:
+            name = mnode[1]
+            info = mnode
+            self.mnode_list[name] = info
+
+        tdSql.query("show dnodes")
+        for dnode in tdSql.queryResult:
+            name = dnode[1]
+            info = dnode
+            self.dnode_list[name] = info
+
+        count = 0
+        is_leader = False
+        mnode_name = ''
+        for k,v in self.mnode_list.items():
+            count +=1
+            # only for 1 mnode
+            mnode_name = k
+
+            if v[2] =='leader':
+                is_leader=True
+
+        if count==1 and is_leader:
+            tdLog.notice("===== depoly cluster success with 1 mnode as leader =====")
+        else:
+            tdLog.exit("===== depoly cluster fail with 1 mnode as leader =====")
+
+        for k ,v in self.dnode_list.items():
+            if k == mnode_name:
+                if v[3]==0:
+                    tdLog.notice("===== depoly cluster mnode only success at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+                else:
+                    tdLog.exit("===== depoly cluster mnode only fail at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+            else:
+                continue
+
+    def create_db_check_vgroups(self):
+
+        tdSql.execute("drop database if exists test")
+        tdSql.execute("create database if not exists test replica 1 duration 300")
+        tdSql.execute("use test")
+        tdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        tdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(5):
+            tdSql.execute("create table sub_tb_{} using stb1 tags({})".format(i,i))
+        tdSql.query("show stables")
+        tdSql.checkRows(1)
+        tdSql.query("show tables")
+        tdSql.checkRows(6)
+
+        tdSql.query("show test.vgroups;")
+        vgroups_infos = {}  # key is id: value is info list
+        for vgroup_info in tdSql.queryResult:
+            vgroup_id = vgroup_info[0]
+            tmp_list = []
+            for role in vgroup_info[3:-4]:
+                if role in ['leader','follower']:
+                    tmp_list.append(role)
+            vgroups_infos[vgroup_id]=tmp_list
+
+        for k , v in vgroups_infos.items():
+            if len(v) ==1 and v[0]=="leader":
+                tdLog.notice(" === create database replica only 1 role leader  check success of vgroup_id {} ======".format(k))
+            else:
+                tdLog.exit(" === create database replica only 1 role leader  check fail of vgroup_id {} ======".format(k))
+
+    def create_db_replica_3_insertdatas(self, dbname, replica_num ,vgroup_nums ,tb_nums , row_nums ):
+        newTdSql=tdCom.newTdSql()
+        drop_db_sql = "drop database if exists {}".format(dbname)
+        create_db_sql = "create database {} replica {} vgroups {}".format(dbname,replica_num,vgroup_nums)
+
+        tdLog.notice(" ==== create database {} and insert rows begin =====".format(dbname))
+        newTdSql.execute(drop_db_sql)
+        newTdSql.execute(create_db_sql)
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        newTdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(tb_nums):
+            sub_tbname = "sub_tb_{}".format(i)
+            newTdSql.execute("create table {} using stb1 tags({})".format(sub_tbname,i))
+            # insert datas about new database
+
+            for row_num in range(row_nums):
+                if row_num %100 ==0:
+                    tdLog.info(" === writing is going now === ")
+                ts = self.ts + 1000*row_num
+                newTdSql.execute(f"insert into {sub_tbname} values ({ts}, {row_num} ,{row_num}, 10 ,1 ,{row_num} ,{row_num},true,'bin_{row_num}','nchar_{row_num}',now) ")
+
+        tdLog.notice(" ==== create database {} and insert rows execute end =====".format(dbname))
+
+    
+
+    def check_insert_status(self, newTdSql ,dbname, tb_nums , row_nums):
+        
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.query("select count(*) from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkData(0 , 0 , tb_nums*row_nums)
+        newTdSql.query("select distinct tbname from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkRows(tb_nums)
+
+    def loop_query_constantly(self, times ,  db_name, tb_nums ,row_nums):
+        newTdSql=tdCom.newTdSql()
+        for loop_time in range(times):
+            tdLog.debug(" === query is going ,this is {}_th query === ".format(loop_time))
+            self.check_insert_status( newTdSql ,db_name, tb_nums , row_nums)
+
+    def run(self): 
+        self.check_setup_cluster_status()
+        self.create_db_check_vgroups()
+        
+        # start writing constantly 
+        writing = threading.Thread(target = self.create_db_replica_3_insertdatas, args=(self.db_name , self.replica , self.vgroups , self.tb_nums , self.row_nums))
+        writing.start()
+        tdSql.query(" show {}.stables ".format(self.db_name))
+        while not tdSql.queryResult:
+            print(tdSql.queryResult)
+            time.sleep(0.1)
+            tdSql.query(" show {}.stables ".format(self.db_name))
+
+        reading = threading.Thread(target = self.loop_query_constantly, args=(self.query_times,self.db_name , self.tb_nums , self.row_nums))
+        reading.start()
+        
+        writing.join()
+        reading.join()
+
+    def stop(self):
+        tdSql.close()
+        tdLog.success(f"{__file__} successfully executed")
+
+tdCases.addLinux(__file__, TDTestCase())
+tdCases.addWindows(__file__, TDTestCase())
\ No newline at end of file

tests/system-test/6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_all_vnode.py

+# author : wenzhouwww
+from ssl import ALERT_DESCRIPTION_CERTIFICATE_UNOBTAINABLE
+import taos
+import sys
+import time
+import os 
+
+from util.log import *
+from util.sql import *
+from util.cases import *
+from util.dnodes import TDDnodes
+from util.dnodes import TDDnode
+from util.cluster import *
+
+import time
+import socket
+import subprocess ,threading
+sys.path.append(os.path.dirname(__file__))
+
+class TDTestCase:
+    def init(self,conn ,logSql):
+        tdLog.debug(f"start to excute {__file__}")
+        tdSql.init(conn.cursor())
+        self.host = socket.gethostname()
+        self.mnode_list = {}
+        self.dnode_list = {}
+        self.ts = 1483200000000
+        self.db_name ='testdb'
+        self.replica = 3 
+        self.vgroups = 1
+        self.tb_nums = 10 
+        self.row_nums = 500
+        self.max_restart_time = 20
+        self.restart_server_times = 10
+        self.dnode_index = 0 
+
+    def getBuildPath(self):
+        selfPath = os.path.dirname(os.path.realpath(__file__))
+        if ("community" in selfPath):
+            projPath = selfPath[:selfPath.find("community")]
+        else:
+            projPath = selfPath[:selfPath.find("tests")]
+
+        for root, dirs, files in os.walk(projPath):
+            if ("taosd" in files):
+                rootRealPath = os.path.dirname(os.path.realpath(root))
+                if ("packaging" not in rootRealPath):
+                    buildPath = root[:len(root) - len("/build/bin")]
+                    break
+        return buildPath
+
+    def check_setup_cluster_status(self):
+        tdSql.query("show mnodes")
+        for mnode in tdSql.queryResult:
+            name = mnode[1]
+            info = mnode
+            self.mnode_list[name] = info
+
+        tdSql.query("show dnodes")
+        for dnode in tdSql.queryResult:
+            name = dnode[1]
+            info = dnode
+            self.dnode_list[name] = info
+
+        count = 0
+        is_leader = False
+        mnode_name = ''
+        for k,v in self.mnode_list.items():
+            count +=1
+            # only for 1 mnode
+            mnode_name = k
+
+            if v[2] =='leader':
+                is_leader=True
+
+        if count==1 and is_leader:
+            tdLog.notice("===== depoly cluster success with 1 mnode as leader =====")
+        else:
+            tdLog.exit("===== depoly cluster fail with 1 mnode as leader =====")
+
+        for k ,v in self.dnode_list.items():
+            if k == mnode_name:
+                if v[3]==0:
+                    tdLog.notice("===== depoly cluster mnode only success at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+                else:
+                    tdLog.exit("===== depoly cluster mnode only fail at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+            else:
+                continue
+
+    def create_db_check_vgroups(self):
+
+        tdSql.execute("drop database if exists test")
+        tdSql.execute("create database if not exists test replica 1 duration 300")
+        tdSql.execute("use test")
+        tdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        tdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(5):
+            tdSql.execute("create table sub_tb_{} using stb1 tags({})".format(i,i))
+        tdSql.query("show stables")
+        tdSql.checkRows(1)
+        tdSql.query("show tables")
+        tdSql.checkRows(6)
+
+        tdSql.query("show test.vgroups;")
+        vgroups_infos = {}  # key is id: value is info list
+
+        for vgroup_info in tdSql.queryResult:
+            vgroup_id = vgroup_info[0]
+            tmp_list = []
+            for role in vgroup_info[3:-4]:
+                if role in ['leader','follower']:
+                    tmp_list.append(role)
+            vgroups_infos[vgroup_id]=tmp_list
+
+        for k , v in vgroups_infos.items():
+            if len(v) ==1 and v[0]=="leader":
+                tdLog.notice(" === create database replica only 1 role leader  check success of vgroup_id {} ======".format(k))
+            else:
+                tdLog.exit(" === create database replica only 1 role leader  check fail of vgroup_id {} ======".format(k))
+
+    def create_db_replica_3_insertdatas(self, dbname, replica_num ,vgroup_nums ,tb_nums , row_nums ):
+
+        newTdSql=tdCom.newTdSql()
+        drop_db_sql = "drop database if exists {}".format(dbname)
+        create_db_sql = "create database {} replica {} vgroups {}".format(dbname,replica_num,vgroup_nums)
+
+        tdLog.notice(" ==== create database {} and insert rows begin =====".format(dbname))
+        newTdSql.execute(drop_db_sql)
+        newTdSql.execute(create_db_sql)
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        newTdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(tb_nums):
+            sub_tbname = "sub_tb_{}".format(i)
+            newTdSql.execute("create table {} using stb1 tags({})".format(sub_tbname,i))
+            # insert datas about new database
+
+            for row_num in range(row_nums):
+                if row_num % (int(row_nums*0.1)) == 0 :
+                    tdLog.notice( " === database {} writing records {} rows".format(dbname , row_num ) )
+                ts = self.ts + 1000*row_num
+
+                newTdSql.execute(f"insert into {sub_tbname} values ({ts}, {row_num} ,{row_num}, 10 ,1 ,{row_num} ,{row_num},true,'bin_{row_num}','nchar_{row_num}',now) ")
+
+        tdLog.notice(" ==== create database {} and insert rows execute end =====".format(dbname))
+
+    
+    def _get_stop_dnode_id(self):
+
+        
+        dnode_lists = list(set(self.dnode_list.keys()) -set(self.mnode_list.keys()))
+        # print(dnode_lists)
+        self.stop_dnode_id = self.dnode_list[dnode_lists[self.dnode_index % 3]][0]
+
+        self.dnode_index += 1 
+
+        return self.stop_dnode_id
+
+    def wait_stop_dnode_OK(self , newTdSql):
+    
+        def _get_status():
+            # newTdSql=tdCom.newTdSql()
+
+            status =  ""
+            newTdSql.query("show dnodes")
+            dnode_infos = newTdSql.queryResult
+            for dnode_info in dnode_infos:
+                id = dnode_info[0]
+                dnode_status = dnode_info[4]
+                if id == self.stop_dnode_id:
+                    status = dnode_status
+                    break
+            return status 
+
+        status = _get_status()
+
+        while status !="offline":
+            time.sleep(0.1)
+            status = _get_status()
+            # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
+        tdLog.notice("==== stop_dnode has stopped , id is {} ====".format(self.stop_dnode_id))
+
+    def wait_start_dnode_OK(self , newTdSql ):
+    
+        def _get_status():
+            # newTdSql=tdCom.newTdSql()
+            status =  ""
+            newTdSql.query("show dnodes")
+            dnode_infos = newTdSql.queryResult
+            for dnode_info in dnode_infos:
+                id = dnode_info[0]
+                dnode_status = dnode_info[4]
+                if id == self.stop_dnode_id:
+                    status = dnode_status
+                    break
+            return status 
+
+        status = _get_status()
+        while status !="ready":
+            time.sleep(0.1)
+            status = _get_status()
+            # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
+        tdLog.notice("==== stop_dnode has restart , id is {} ====".format(self.stop_dnode_id))
+
+    def get_leader_infos(self , newTdSql , dbname):
+        
+        # newTdSql=tdCom.newTdSql()
+        newTdSql.query("show {}.vgroups".format(dbname))
+        vgroup_infos = newTdSql.queryResult
+
+        leader_infos = set()
+        for vgroup_info in vgroup_infos:
+            leader_infos.add(vgroup_info[3:-4])
+
+        return leader_infos
+
+    def check_revote_leader_success(self, dbname, before_leader_infos , after_leader_infos):
+
+        check_status = False
+        vote_act = set(set(after_leader_infos)-set(before_leader_infos))
+        if not vote_act:
+            print("=======before_revote_leader_infos ======\n" , before_leader_infos)
+            print("=======after_revote_leader_infos ======\n" , after_leader_infos)
+            tdLog.exit(" ===maybe revote not occured , there is no dnode offline ====")
+        else:
+            for vgroup_info in vote_act:
+                for ind , role in enumerate(vgroup_info):  
+                    if role==self.stop_dnode_id:
+
+                        if vgroup_info[ind+1] =="offline" and "leader" in vgroup_info:
+                            tdLog.notice(" === revote leader ok , leader is {} now   ====".format(vgroup_info[list(vgroup_info).index("leader")-1]))
+                            check_status = True
+                        elif vgroup_info[ind+1] !="offline":
+                            tdLog.notice(" === dnode {} should be offline ".format(self.stop_dnode_id))
+                        else:
+                            continue
+                        break
+        return check_status
+
+
+    def check_insert_status(self, newTdSql , dbname, tb_nums , row_nums):
+
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.query("select count(*) from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkData(0 , 0 , tb_nums*row_nums)
+        newTdSql.query("select distinct tbname from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkRows(tb_nums)
+
+    def loop_query_constantly(self, times ,  db_name, tb_nums ,row_nums):
+
+        newTdSql=tdCom.newTdSql()
+        for loop_time in range(times):
+            tdLog.debug(" === query is going ,this is {}_th query === ".format(loop_time))
+            
+            self.check_insert_status( newTdSql ,db_name, tb_nums , row_nums)
+
+    
+    def loop_restart_follower_constantly(self, times , db_name):
+
+        tdDnodes = cluster.dnodes
+        newTdSql=tdCom.newTdSql()
+
+        for loop_time in range(times):
+
+            self.stop_dnode_id = self._get_stop_dnode_id()
+            
+            # print(self.stop_dnode_id)
+            # begin stop dnode 
+
+            # before_leader_infos = self.get_leader_infos( newTdSql ,db_name)
+            tdDnodes[self.stop_dnode_id-1].stoptaosd()
+            self.wait_stop_dnode_OK(newTdSql)
+
+            # start = time.time()
+            # # get leader info after stop 
+            # after_leader_infos = self.get_leader_infos(newTdSql , db_name)
+             
+            # revote_status = self.check_revote_leader_success(db_name ,before_leader_infos , after_leader_infos)
+
+            # # append rows of stablename when dnode stop make sure revote leaders
+            
+            # while not revote_status:
+            #     after_leader_infos = self.get_leader_infos(newTdSql , db_name)
+            #     revote_status = self.check_revote_leader_success(db_name ,before_leader_infos , after_leader_infos)
+
+            # end = time.time()
+            # time_cost = end - start 
+            # tdLog.notice(" ==== revote leader of database {} cost time {}  ====".format(db_name , time_cost))
+
+            tdLog.notice(" === this is {}_th restart taosd === ".format(loop_time))
+
+            # begin start dnode 
+            start = time.time()
+            tdDnodes[self.stop_dnode_id-1].starttaosd()
+            self.wait_start_dnode_OK(newTdSql)
+            time.sleep(5)
+            end = time.time()
+            time_cost = int(end -start)
+            if time_cost > self.max_restart_time:
+                tdLog.exit(" ==== restart dnode {} cost too much time , please check ====".format(self.stop_dnode_id))
+            
+
+
+    def run(self): 
+
+        self.check_setup_cluster_status()
+        self.create_db_check_vgroups()
+        
+        # start writing constantly 
+        writing = threading.Thread(target = self.create_db_replica_3_insertdatas, args=(self.db_name , self.replica , self.vgroups , self.tb_nums , self.row_nums))
+        writing.start()
+        tdSql.query(" show {}.stables ".format(self.db_name))
+        while not tdSql.queryResult:
+            print(tdSql.queryResult)
+            time.sleep(0.1)
+            tdSql.query(" show {}.stables ".format(self.db_name))
+
+        restart_servers = threading.Thread(target = self.loop_restart_follower_constantly, args = (self.restart_server_times ,self.db_name))
+        restart_servers.start()
+
+        # reading = threading.Thread(target = self.loop_query_constantly, args=(1000,self.db_name , self.tb_nums , self.row_nums))
+        # reading.start()
+        
+        writing.join()
+        # reading.join()
+        restart_servers.join()
+
+    def stop(self):
+        tdSql.close()
+        tdLog.success(f"{__file__} successfully executed")
+
+tdCases.addLinux(__file__, TDTestCase())
+tdCases.addWindows(__file__, TDTestCase())
\ No newline at end of file

tests/system-test/6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_follower.py

+# author : wenzhouwww
+from ssl import ALERT_DESCRIPTION_CERTIFICATE_UNOBTAINABLE
+import taos
+import sys
+import time
+import os 
+
+from util.log import *
+from util.sql import *
+from util.cases import *
+from util.dnodes import TDDnodes
+from util.dnodes import TDDnode
+from util.cluster import *
+
+import time
+import socket
+import subprocess ,threading
+sys.path.append(os.path.dirname(__file__))
+
+class TDTestCase:
+    def init(self,conn ,logSql):
+        tdLog.debug(f"start to excute {__file__}")
+        tdSql.init(conn.cursor())
+        self.host = socket.gethostname()
+        self.mnode_list = {}
+        self.dnode_list = {}
+        self.ts = 1483200000000
+        self.db_name ='testdb'
+        self.replica = 3 
+        self.vgroups = 10
+        self.tb_nums = 10 
+        self.row_nums = 100
+        self.max_restart_time = 20
+        self.restart_server_times = 5
+        self.query_times = 100 
+
+    def getBuildPath(self):
+        selfPath = os.path.dirname(os.path.realpath(__file__))
+        if ("community" in selfPath):
+            projPath = selfPath[:selfPath.find("community")]
+        else:
+            projPath = selfPath[:selfPath.find("tests")]
+
+        for root, dirs, files in os.walk(projPath):
+            if ("taosd" in files):
+                rootRealPath = os.path.dirname(os.path.realpath(root))
+                if ("packaging" not in rootRealPath):
+                    buildPath = root[:len(root) - len("/build/bin")]
+                    break
+        return buildPath
+
+    def check_setup_cluster_status(self):
+        tdSql.query("show mnodes")
+        for mnode in tdSql.queryResult:
+            name = mnode[1]
+            info = mnode
+            self.mnode_list[name] = info
+
+        tdSql.query("show dnodes")
+        for dnode in tdSql.queryResult:
+            name = dnode[1]
+            info = dnode
+            self.dnode_list[name] = info
+
+        count = 0
+        is_leader = False
+        mnode_name = ''
+        for k,v in self.mnode_list.items():
+            count +=1
+            # only for 1 mnode
+            mnode_name = k
+
+            if v[2] =='leader':
+                is_leader=True
+
+        if count==1 and is_leader:
+            tdLog.notice("===== depoly cluster success with 1 mnode as leader =====")
+        else:
+            tdLog.exit("===== depoly cluster fail with 1 mnode as leader =====")
+
+        for k ,v in self.dnode_list.items():
+            if k == mnode_name:
+                if v[3]==0:
+                    tdLog.notice("===== depoly cluster mnode only success at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+                else:
+                    tdLog.exit("===== depoly cluster mnode only fail at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+            else:
+                continue
+
+    def create_db_check_vgroups(self):
+
+        tdSql.execute("drop database if exists test")
+        tdSql.execute("create database if not exists test replica 1 duration 300")
+        tdSql.execute("use test")
+        tdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        tdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(5):
+            tdSql.execute("create table sub_tb_{} using stb1 tags({})".format(i,i))
+        tdSql.query("show stables")
+        tdSql.checkRows(1)
+        tdSql.query("show tables")
+        tdSql.checkRows(6)
+
+        tdSql.query("show test.vgroups;")
+        vgroups_infos = {}  # key is id: value is info list
+        for vgroup_info in tdSql.queryResult:
+            vgroup_id = vgroup_info[0]
+            tmp_list = []
+            for role in vgroup_info[3:-4]:
+                if role in ['leader','follower']:
+                    tmp_list.append(role)
+            vgroups_infos[vgroup_id]=tmp_list
+
+        for k , v in vgroups_infos.items():
+            if len(v) ==1 and v[0]=="leader":
+                tdLog.notice(" === create database replica only 1 role leader  check success of vgroup_id {} ======".format(k))
+            else:
+                tdLog.exit(" === create database replica only 1 role leader  check fail of vgroup_id {} ======".format(k))
+
+    def create_db_replica_3_insertdatas(self, dbname, replica_num ,vgroup_nums ,tb_nums , row_nums ):
+        newTdSql=tdCom.newTdSql()
+        drop_db_sql = "drop database if exists {}".format(dbname)
+        create_db_sql = "create database {} replica {} vgroups {}".format(dbname,replica_num,vgroup_nums)
+
+        tdLog.notice(" ==== create database {} and insert rows begin =====".format(dbname))
+        newTdSql.execute(drop_db_sql)
+        newTdSql.execute(create_db_sql)
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        newTdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(tb_nums):
+            sub_tbname = "sub_tb_{}".format(i)
+            newTdSql.execute("create table {} using stb1 tags({})".format(sub_tbname,i))
+            # insert datas about new database
+
+            for row_num in range(row_nums):
+                if row_num % (int(row_nums*0.1)) == 0 :
+                    tdLog.notice( " === database {} writing records {} rows".format(dbname , row_num ) )
+                ts = self.ts + 1000*row_num
+                newTdSql.execute(f"insert into {sub_tbname} values ({ts}, {row_num} ,{row_num}, 10 ,1 ,{row_num} ,{row_num},true,'bin_{row_num}','nchar_{row_num}',now) ")
+
+        tdLog.notice(" ==== create database {} and insert rows execute end =====".format(dbname))
+
+    
+    def _get_stop_dnode_id(self,dbname ,dnode_role):
+        tdSql.query("show {}.vgroups".format(dbname))
+        vgroup_infos = tdSql.queryResult
+        status = False
+        for vgroup_info in vgroup_infos:
+            if "error" not in vgroup_info:
+                status = True
+            else:
+                status = False
+        while status!=True :
+            time.sleep(0.1)
+            tdSql.query("show {}.vgroups".format(dbname))
+            vgroup_infos = tdSql.queryResult
+            for vgroup_info in vgroup_infos:
+                if "error" not in vgroup_info:
+                    status = True
+                else:
+                    status = False
+            # print(status)
+        for vgroup_info in vgroup_infos:
+            leader_infos = vgroup_info[3:-4] 
+            # print(vgroup_info)
+            for ind ,role in enumerate(leader_infos):
+                if role == dnode_role:
+                    # print(ind,leader_infos)
+                    self.stop_dnode_id = leader_infos[ind-1]
+                    break
+
+        return self.stop_dnode_id
+
+    def wait_stop_dnode_OK(self , newTdSql):
+    
+        def _get_status():
+            # newTdSql=tdCom.newTdSql()
+
+            status =  ""
+            newTdSql.query("show dnodes")
+            dnode_infos = newTdSql.queryResult
+            for dnode_info in dnode_infos:
+                id = dnode_info[0]
+                dnode_status = dnode_info[4]
+                if id == self.stop_dnode_id:
+                    status = dnode_status
+                    break
+            return status 
+
+        status = _get_status()
+
+        while status !="offline":
+            time.sleep(0.1)
+            status = _get_status()
+            # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
+        tdLog.notice("==== stop_dnode has stopped , id is {} ====".format(self.stop_dnode_id))
+
+    def wait_start_dnode_OK(self , newTdSql ):
+    
+        def _get_status():
+            # newTdSql=tdCom.newTdSql()
+            status =  ""
+            newTdSql.query("show dnodes")
+            dnode_infos = newTdSql.queryResult
+            for dnode_info in dnode_infos:
+                id = dnode_info[0]
+                dnode_status = dnode_info[4]
+                if id == self.stop_dnode_id:
+                    status = dnode_status
+                    break
+            return status 
+
+        status = _get_status()
+        while status !="ready":
+            time.sleep(0.1)
+            status = _get_status()
+            # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
+        tdLog.notice("==== stop_dnode has restart , id is {} ====".format(self.stop_dnode_id))
+
+    def check_insert_status(self, newTdSql , dbname, tb_nums , row_nums):
+
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.query("select count(*) from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkData(0 , 0 , tb_nums*row_nums)
+        newTdSql.query("select distinct tbname from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkRows(tb_nums)
+
+    def loop_query_constantly(self, times ,  db_name, tb_nums ,row_nums):
+
+        newTdSql=tdCom.newTdSql()
+        for loop_time in range(times):
+            tdLog.debug(" === query is going ,this is {}_th query === ".format(loop_time))
+            
+            self.check_insert_status( newTdSql ,db_name, tb_nums , row_nums)
+
+    
+    def loop_restart_follower_constantly(self, times , db_name):
+
+        tdDnodes = cluster.dnodes
+        newTdSql=tdCom.newTdSql()
+
+        for loop_time in range(times):
+
+            self.stop_dnode_id = self._get_stop_dnode_id(db_name , "follower")
+            
+            print(self.stop_dnode_id)
+            # begin stop dnode 
+            start = time.time()
+            tdDnodes[self.stop_dnode_id-1].stoptaosd()
+            self.wait_stop_dnode_OK(newTdSql)
+
+            tdLog.notice(" === this is {}_th restart taosd === ".format(loop_time))
+
+            # begin start dnode 
+            tdDnodes[self.stop_dnode_id-1].starttaosd()
+            self.wait_start_dnode_OK(newTdSql)
+            end = time.time()
+            time_cost = int(end -start)
+            if time_cost > self.max_restart_time:
+                tdLog.exit(" ==== restart dnode {} cost too much time , please check ====".format(self.stop_dnode_id))
+            
+
+
+    def run(self): 
+
+        self.check_setup_cluster_status()
+        self.create_db_check_vgroups()
+        
+        # start writing constantly 
+        writing = threading.Thread(target = self.create_db_replica_3_insertdatas, args=(self.db_name , self.replica , self.vgroups , self.tb_nums , self.row_nums))
+        writing.start()
+        tdSql.query(" show {}.stables ".format(self.db_name))
+        while not tdSql.queryResult:
+            print(tdSql.queryResult)
+            time.sleep(0.1)
+            tdSql.query(" show {}.stables ".format(self.db_name))
+
+        restart_servers = threading.Thread(target = self.loop_restart_follower_constantly, args = (self.restart_server_times ,self.db_name))
+        restart_servers.start()
+
+        reading = threading.Thread(target = self.loop_query_constantly, args=(self.query_times,self.db_name , self.tb_nums , self.row_nums))
+        reading.start()
+        
+        writing.join()
+        reading.join()
+        restart_servers.join()
+
+    def stop(self):
+        tdSql.close()
+        tdLog.success(f"{__file__} successfully executed")
+
+tdCases.addLinux(__file__, TDTestCase())
+tdCases.addWindows(__file__, TDTestCase())
\ No newline at end of file

tests/system-test/6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_leader.py

+# author : wenzhouwww
+from ssl import ALERT_DESCRIPTION_CERTIFICATE_UNOBTAINABLE
+import taos
+import sys
+import time
+import os 
+
+from util.log import *
+from util.sql import *
+from util.cases import *
+from util.dnodes import TDDnodes
+from util.dnodes import TDDnode
+from util.cluster import *
+
+import time
+import socket
+import subprocess ,threading
+sys.path.append(os.path.dirname(__file__))
+
+class TDTestCase:
+    def init(self,conn ,logSql):
+        tdLog.debug(f"start to excute {__file__}")
+        tdSql.init(conn.cursor())
+        self.host = socket.gethostname()
+        self.mnode_list = {}
+        self.dnode_list = {}
+        self.ts = 1483200000000
+        self.db_name ='testdb'
+        self.replica = 3 
+        self.vgroups = 10
+        self.tb_nums = 10 
+        self.row_nums = 100
+        self.max_restart_time = 20
+        self.restart_server_times = 10 
+        self.query_times = 100
+
+    def getBuildPath(self):
+        selfPath = os.path.dirname(os.path.realpath(__file__))
+        if ("community" in selfPath):
+            projPath = selfPath[:selfPath.find("community")]
+        else:
+            projPath = selfPath[:selfPath.find("tests")]
+
+        for root, dirs, files in os.walk(projPath):
+            if ("taosd" in files):
+                rootRealPath = os.path.dirname(os.path.realpath(root))
+                if ("packaging" not in rootRealPath):
+                    buildPath = root[:len(root) - len("/build/bin")]
+                    break
+        return buildPath
+
+    def check_setup_cluster_status(self):
+        tdSql.query("show mnodes")
+        for mnode in tdSql.queryResult:
+            name = mnode[1]
+            info = mnode
+            self.mnode_list[name] = info
+
+        tdSql.query("show dnodes")
+        for dnode in tdSql.queryResult:
+            name = dnode[1]
+            info = dnode
+            self.dnode_list[name] = info
+
+        count = 0
+        is_leader = False
+        mnode_name = ''
+        for k,v in self.mnode_list.items():
+            count +=1
+            # only for 1 mnode
+            mnode_name = k
+
+            if v[2] =='leader':
+                is_leader=True
+
+        if count==1 and is_leader:
+            tdLog.notice("===== depoly cluster success with 1 mnode as leader =====")
+        else:
+            tdLog.exit("===== depoly cluster fail with 1 mnode as leader =====")
+
+        for k ,v in self.dnode_list.items():
+            if k == mnode_name:
+                if v[3]==0:
+                    tdLog.notice("===== depoly cluster mnode only success at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+                else:
+                    tdLog.exit("===== depoly cluster mnode only fail at {} , support_vnodes is {} ".format(mnode_name,v[3]))
+            else:
+                continue
+
+    def create_db_check_vgroups(self):
+
+        tdSql.execute("drop database if exists test")
+        tdSql.execute("create database if not exists test replica 1 duration 300")
+        tdSql.execute("use test")
+        tdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        tdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(5):
+            tdSql.execute("create table sub_tb_{} using stb1 tags({})".format(i,i))
+        tdSql.query("show stables")
+        tdSql.checkRows(1)
+        tdSql.query("show tables")
+        tdSql.checkRows(6)
+
+        tdSql.query("show test.vgroups;")
+        vgroups_infos = {}  # key is id: value is info list
+        for vgroup_info in tdSql.queryResult:
+            vgroup_id = vgroup_info[0]
+            tmp_list = []
+            for role in vgroup_info[3:-4]:
+                if role in ['leader','follower']:
+                    tmp_list.append(role)
+            vgroups_infos[vgroup_id]=tmp_list
+
+        for k , v in vgroups_infos.items():
+            if len(v) ==1 and v[0]=="leader":
+                tdLog.notice(" === create database replica only 1 role leader  check success of vgroup_id {} ======".format(k))
+            else:
+                tdLog.exit(" === create database replica only 1 role leader  check fail of vgroup_id {} ======".format(k))
+
+    def create_db_replica_3_insertdatas(self, dbname, replica_num ,vgroup_nums ,tb_nums , row_nums ):
+        newTdSql=tdCom.newTdSql()
+        drop_db_sql = "drop database if exists {}".format(dbname)
+        create_db_sql = "create database {} replica {} vgroups {}".format(dbname,replica_num,vgroup_nums)
+
+        tdLog.notice(" ==== create database {} and insert rows begin =====".format(dbname))
+        newTdSql.execute(drop_db_sql)
+        newTdSql.execute(create_db_sql)
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.execute(
+        '''create table stb1
+        (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+        tags (t1 int)
+        '''
+        )
+        newTdSql.execute(
+            '''
+            create table t1
+            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(32),c9 nchar(32), c10 timestamp)
+            '''
+        )
+        
+        for i in range(tb_nums):
+            sub_tbname = "sub_tb_{}".format(i)
+            newTdSql.execute("create table {} using stb1 tags({})".format(sub_tbname,i))
+            # insert datas about new database
+
+            for row_num in range(row_nums):
+                if row_num % (int(row_nums*0.1)) == 0 :
+                    tdLog.notice( " === database {} writing records {} rows".format(dbname , row_num ) )
+                ts = self.ts + 1000*row_num
+                newTdSql.execute(f"insert into {sub_tbname} values ({ts}, {row_num} ,{row_num}, 10 ,1 ,{row_num} ,{row_num},true,'bin_{row_num}','nchar_{row_num}',now) ")
+
+        tdLog.notice(" ==== create database {} and insert rows execute end =====".format(dbname))
+
+    def _get_stop_dnode_id(self,dbname ,dnode_role):
+        tdSql.query("show {}.vgroups".format(dbname))
+        vgroup_infos = tdSql.queryResult
+        status = False
+        for vgroup_info in vgroup_infos:
+            if "error" not in vgroup_info:
+                status = True
+            else:
+                status = False
+        while status!=True :
+            time.sleep(0.1)
+            tdSql.query("show {}.vgroups".format(dbname))
+            vgroup_infos = tdSql.queryResult
+            for vgroup_info in vgroup_infos:
+                if "error" not in vgroup_info:
+                    status = True
+                else:
+                    status = False
+            # print(status)
+        for vgroup_info in vgroup_infos:
+            leader_infos = vgroup_info[3:-4] 
+            # print(vgroup_info)
+            for ind ,role in enumerate(leader_infos):
+                if role == dnode_role:
+                    # print(ind,leader_infos)
+                    self.stop_dnode_id = leader_infos[ind-1]
+                    break
+
+        return self.stop_dnode_id
+
+    def wait_stop_dnode_OK(self , newTdSql):
+    
+        def _get_status():
+            # newTdSql=tdCom.newTdSql()
+
+            status =  ""
+            newTdSql.query("show dnodes")
+            dnode_infos = newTdSql.queryResult
+            for dnode_info in dnode_infos:
+                id = dnode_info[0]
+                dnode_status = dnode_info[4]
+                if id == self.stop_dnode_id:
+                    status = dnode_status
+                    break
+            return status 
+
+        status = _get_status()
+
+        while status !="offline":
+            time.sleep(0.1)
+            status = _get_status()
+            # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
+        tdLog.notice("==== stop_dnode has stopped , id is {} ====".format(self.stop_dnode_id))
+
+    def wait_start_dnode_OK(self , newTdSql ):
+    
+        def _get_status():
+            # newTdSql=tdCom.newTdSql()
+            status =  ""
+            newTdSql.query("show dnodes")
+            dnode_infos = newTdSql.queryResult
+            for dnode_info in dnode_infos:
+                id = dnode_info[0]
+                dnode_status = dnode_info[4]
+                if id == self.stop_dnode_id:
+                    status = dnode_status
+                    break
+            return status 
+
+        status = _get_status()
+        while status !="ready":
+            time.sleep(0.1)
+            status = _get_status()
+            # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
+        tdLog.notice("==== stop_dnode has restart , id is {} ====".format(self.stop_dnode_id))
+
+    def get_leader_infos(self , newTdSql , dbname):
+        
+        # newTdSql=tdCom.newTdSql()
+        newTdSql.query("show {}.vgroups".format(dbname))
+        vgroup_infos = newTdSql.queryResult
+
+        leader_infos = set()
+        for vgroup_info in vgroup_infos:
+            leader_infos.add(vgroup_info[3:-4])
+
+        return leader_infos
+
+    def check_revote_leader_success(self, dbname, before_leader_infos , after_leader_infos):
+
+        check_status = False
+        vote_act = set(set(after_leader_infos)-set(before_leader_infos))
+        if not vote_act:
+            print("=======before_revote_leader_infos ======\n" , before_leader_infos)
+            print("=======after_revote_leader_infos ======\n" , after_leader_infos)
+            tdLog.exit(" ===maybe revote not occured , there is no dnode offline ====")
+        else:
+            for vgroup_info in vote_act:
+                for ind , role in enumerate(vgroup_info):  
+                    if role==self.stop_dnode_id:
+
+                        if vgroup_info[ind+1] =="offline" and "leader" in vgroup_info:
+                            tdLog.notice(" === revote leader ok , leader is {} now   ====".format(vgroup_info[list(vgroup_info).index("leader")-1]))
+                            check_status = True
+                        elif vgroup_info[ind+1] !="offline":
+                            tdLog.notice(" === dnode {} should be offline ".format(self.stop_dnode_id))
+                        else:
+                            continue
+                        break
+        return check_status
+
+
+    def check_insert_status(self, newTdSql , dbname, tb_nums , row_nums):
+
+        newTdSql.execute("use {}".format(dbname))
+        newTdSql.query("select count(*) from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkData(0 , 0 , tb_nums*row_nums)
+        newTdSql.query("select distinct tbname from {}.{}".format(dbname,'stb1'))
+        # tdSql.checkRows(tb_nums)
+
+    def loop_query_constantly(self, times ,  db_name, tb_nums ,row_nums):
+
+        newTdSql=tdCom.newTdSql()
+        for loop_time in range(times):
+            tdLog.debug(" === query is going ,this is {}_th query === ".format(loop_time))
+            
+            self.check_insert_status( newTdSql ,db_name, tb_nums , row_nums)
+
+    
+    def loop_restart_follower_constantly(self, times , db_name):
+
+        tdDnodes = cluster.dnodes
+        newTdSql=tdCom.newTdSql()
+
+        for loop_time in range(times):
+
+            self.stop_dnode_id = self._get_stop_dnode_id(db_name , "leader")
+            
+            # print(self.stop_dnode_id)
+            # begin stop dnode 
+            start = time.time()
+
+            before_leader_infos = self.get_leader_infos( newTdSql ,db_name)
+            tdDnodes[self.stop_dnode_id-1].stoptaosd()
+            self.wait_stop_dnode_OK(newTdSql)
+
+            start = time.time()
+            # get leader info after stop 
+            after_leader_infos = self.get_leader_infos(newTdSql , db_name)
+             
+            revote_status = self.check_revote_leader_success(db_name ,before_leader_infos , after_leader_infos)
+
+            # append rows of stablename when dnode stop make sure revote leaders
+            
+            while not revote_status:
+                after_leader_infos = self.get_leader_infos(newTdSql , db_name)
+                revote_status = self.check_revote_leader_success(db_name ,before_leader_infos , after_leader_infos)
+
+            end = time.time()
+            time_cost = end - start 
+            tdLog.notice(" ==== revote leader of database {} cost time {}  ====".format(db_name , time_cost))
+
+            tdLog.notice(" === this is {}_th restart taosd === ".format(loop_time))
+
+            # begin start dnode 
+            tdDnodes[self.stop_dnode_id-1].starttaosd()
+            self.wait_start_dnode_OK(newTdSql)
+            end = time.time()
+            time_cost = int(end -start)
+            if time_cost > self.max_restart_time:
+                tdLog.exit(" ==== restart dnode {} cost too much time , please check ====".format(self.stop_dnode_id))
+            
+
+
+    def run(self): 
+
+        self.check_setup_cluster_status()
+        self.create_db_check_vgroups()
+        
+        # start writing constantly 
+        writing = threading.Thread(target = self.create_db_replica_3_insertdatas, args=(self.db_name , self.replica , self.vgroups , self.tb_nums , self.row_nums))
+        writing.start()
+        tdSql.query(" show {}.stables ".format(self.db_name))
+        while not tdSql.queryResult:
+            print(tdSql.queryResult)
+            time.sleep(0.1)
+            tdSql.query(" show {}.stables ".format(self.db_name))
+
+        restart_servers = threading.Thread(target = self.loop_restart_follower_constantly, args = (self.restart_server_times ,self.db_name))
+        restart_servers.start()
+
+        reading = threading.Thread(target = self.loop_query_constantly, args=(self.query_times,self.db_name , self.tb_nums , self.row_nums))
+        reading.start()
+        
+        writing.join()
+        reading.join()
+        restart_servers.join()
+
+    def stop(self):
+        tdSql.close()
+        tdLog.success(f"{__file__} successfully executed")
+
+tdCases.addLinux(__file__, TDTestCase())
+tdCases.addWindows(__file__, TDTestCase())
\ No newline at end of file

tests/system-test/6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_follower_sync.py

@@ -220,25 +220,43 @@ class TDTestCase:
             status_OK = self.mycheckRows("select distinct tbname from {}.{}".format(dbname,stablename) ,tb_nums)
             tdLog.debug(" ==== check insert tbnames first failed , this is {}_th retry check tbnames of database {}".format(count , dbname))
             count += 1
-    def _get_stop_dnode_id(self,dbname):
+
+    
+    def _get_stop_dnode_id(self,dbname ,dnode_role):
         tdSql.query("show {}.vgroups".format(dbname))
         vgroup_infos = tdSql.queryResult
+        status = False
+        for vgroup_info in vgroup_infos:
+            if "error" not in vgroup_info:
+                status = True
+            else:
+                status = False
+        while status!=True :
+            time.sleep(0.1)
+            tdSql.query("show {}.vgroups".format(dbname))
+            vgroup_infos = tdSql.queryResult
+            for vgroup_info in vgroup_infos:
+                if "error" not in vgroup_info:
+                    status = True
+                else:
+                    status = False
+            # print(status)
         for vgroup_info in vgroup_infos:
             leader_infos = vgroup_info[3:-4]
             # print(vgroup_info)
             for ind ,role in enumerate(leader_infos):
-                if role =='follower':
+                if role == dnode_role:
                     # print(ind,leader_infos)
                     self.stop_dnode_id = leader_infos[ind-1]
                     break
 
-
         return self.stop_dnode_id
 
-    def wait_stop_dnode_OK(self):
+    def wait_stop_dnode_OK(self ,newTdSql):
+    
 
         def _get_status():
-            newTdSql=tdCom.newTdSql()
+            # newTdSql=tdCom.newTdSql()
 
             status =  ""
             newTdSql.query("show dnodes")
@@ -258,10 +276,11 @@ class TDTestCase:
             # tdLog.notice("==== stop dnode has not been stopped , endpoint is {}".format(self.stop_dnode))
         tdLog.notice("==== stop_dnode has stopped , id is {} ====".format(self.stop_dnode_id))
 
-    def wait_start_dnode_OK(self):
+
+    def wait_start_dnode_OK(self ,newTdSql):
     
         def _get_status():
-            newTdSql=tdCom.newTdSql()
+            # newTdSql=tdCom.newTdSql()
             status =  ""
             newTdSql.query("show dnodes")
             dnode_infos = newTdSql.queryResult
@@ -369,12 +388,15 @@ class TDTestCase:
     def sync_run_case(self):
         # stop follower and insert datas , update tables and create new stables
         tdDnodes=cluster.dnodes
+        newTdSql =  tdCom.newTdSql()
         for loop in range(self.loop_restart_times):
             db_name = "sync_db_{}".format(loop)
             stablename = 'stable_{}'.format(loop)
             self.create_database(dbname = db_name ,replica_num= self.replica  , vgroup_nums= 1)
             self.create_stable_insert_datas(dbname = db_name , stablename = stablename , tb_nums= 10 ,row_nums= 10 )
-            self.stop_dnode_id = self._get_stop_dnode_id(db_name)
+
+            self.stop_dnode_id = self._get_stop_dnode_id(db_name , "follower")
+            
 
             # check rows of datas
 
@@ -384,7 +406,9 @@ class TDTestCase:
             start = time.time()
             tdDnodes[self.stop_dnode_id-1].stoptaosd()
 
-            self.wait_stop_dnode_OK()
+        
+            self.wait_stop_dnode_OK(newTdSql)
+
 
             # append rows of stablename when dnode stop
 
@@ -402,7 +426,7 @@ class TDTestCase:
 
             # begin start dnode
             tdDnodes[self.stop_dnode_id-1].starttaosd()
-            self.wait_start_dnode_OK()
+            self.wait_start_dnode_OK(newTdSql)
             end = time.time()
             time_cost = int(end -start)
             if time_cost > self.max_restart_time:
@@ -419,7 +443,7 @@ class TDTestCase:
         def _restart_dnode_of_db_unsync(dbname):
             start = time.time()
             tdDnodes=cluster.dnodes
-            self.stop_dnode_id = self._get_stop_dnode_id(dbname)
+            self.stop_dnode_id = self._get_stop_dnode_id(dbname ,"follower")
             # begin restart dnode
             tdDnodes[self.stop_dnode_id-1].stoptaosd()
             self.wait_stop_dnode_OK()

tests/system-test/clusterCase.sh

@@ -19,4 +19,28 @@ python3 ./test.py -f 6-cluster/5dnode3mnodeSep1VnodeStopMnodeCreateStb.py  -N 5
 # python3 ./test.py -f 6-cluster/5dnode3mnodeRestartDnodeInsertData.py -N 5 -M 3
  python3 ./test.py -f 6-cluster/5dnode3mnodeAdd1Ddnoe.py -N 6 -M 3 -C 5 
 
+# test case of vnode 
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_createDb_replica1.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica1_insertdatas.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_vgroups.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_vgroups_stopOne.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_follower_sync.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_follower_unsync.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_follower_unsync_force_stop.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_leader.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_leader_forece_stop.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_querydatas_stop_follower.py  -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_querydatas_stop_follower_force_stop.py  -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_querydatas_stop_leader.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_querydatas_stop_leader_force_stop.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica1_insertdatas.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica1_insertdatas_querys.py -N 4 -M 1 
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys.py -N 4 -M 1 
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_follower.py -N 4 -M 1 
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_leader.py -N 4 -M 1
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_all_vnode.py   -N 4 -M 1 
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_all_dnodes.py -N 4 -M 1 
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_force_stop_all_dnodes.py -N 4 -M 1 
+python3 ./test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_mnode3_insertdatas_querys.py -N 4 -M 1
 

tests/system-test/fulltest.sh

@@ -193,6 +193,30 @@ python3 ./test.py -f 6-cluster/5dnode3mnodeRecreateMnode.py  -N 5 -M 3
 python3 ./test.py -f 6-cluster/5dnode3mnodeStopFollowerLeader.py  -N 5 -M 3 
 python3 ./test.py -f 6-cluster/5dnode3mnodeStop2Follower.py  -N 5 -M 3 
 
+# vnode case 
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_createDb_replica1.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica1_insertdatas.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica1_insertdatas_querys.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_force_stop_all_dnodes.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_all_vnode.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_follower.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys_loop_restart_leader.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_querys.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_all_dnodes.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_follower_sync.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_follower_unsync_force_stop.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_follower_unsync.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_leader_forece_stop.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_insertdatas_stop_leader.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_mnode3_insertdatas_querys.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_querydatas_stop_follower_force_stop.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_querydatas_stop_follower.py -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_querydatas_stop_leader_force_stop.py -N 4 -M 1 
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_querydatas_stop_leader.py -N 4 -M 1 
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_vgroups.py  -N 4 -M 1
+# python3 test.py -f 6-cluster/vnode/4dnode1mnode_basic_replica3_vgroups_stopOne.py -N 4 -M 1
+
 
 python3 ./test.py -f 7-tmq/dropDbR3ConflictTransaction.py -N 3
 python3 ./test.py -f 7-tmq/basic5.py