Merge pull request #17143 from taosdata/3.0

release: build 3.0.1.4

cmake/taostools_CMakeLists.txt.in

@@ -2,7 +2,7 @@
 # taos-tools
 ExternalProject_Add(taos-tools
         GIT_REPOSITORY https://github.com/taosdata/taos-tools.git
-        GIT_TAG f03c09a
+        GIT_TAG 70f5a1c
         SOURCE_DIR "${TD_SOURCE_DIR}/tools/taos-tools"
         BINARY_DIR ""
         #BUILD_IN_SOURCE TRUE

docs/en/07-develop/03-insert-data/02-influxdb-line.mdx

@@ -37,7 +37,8 @@ meters,location=California.LosAngeles,groupid=2 current=13.4,voltage=223,phase=0
 - All the data in `tag_set` will be converted to nchar type automatically .
 - Each data in `field_set` must be self-descriptive for its data type. For example 1.2f32 means a value 1.2 of float type. Without the "f" type suffix, it will be treated as type double.
 - Multiple kinds of precision can be used for the `timestamp` field. Time precision can be from nanosecond (ns) to hour (h).
-
+- You can configure smlChildTableName in taos.cfg to specify table names, for example, `smlChildTableName=tname`. You can insert `st,tname=cpul,t1=4 c1=3 1626006833639000000` and the cpu1 table will be automatically created. Note that if multiple rows have the same tname but different tag_set values, the tag_set of the first row is used to create the table and the others are ignored.
+- It is assumed that the order of field_set in a supertable is consistent, meaning that the first record contains all fields and subsequent records store fields in the same order. If the order is not consistent, set smlDataFormat in taos.cfg to false. Otherwise, data will be written out of order and a database error will occur.(smlDataFormat in taos.cfg default to false after version of 3.0.1.3)
 :::
 
 For more details please refer to [InfluxDB Line Protocol](https://docs.influxdata.com/influxdb/v2.0/reference/syntax/line-protocol/) and [TDengine Schemaless](/reference/schemaless/#Schemaless-Line-Protocol)
@@ -64,3 +65,7 @@ For more details please refer to [InfluxDB Line Protocol](https://docs.influxdat
     <CLine />
   </TabItem>
 </Tabs>
+
+## Query Examples
+If you want query the data of `location=California.LosAngeles,groupid=2`，here is the query sql：
+select * from `meters.voltage` where location="California.LosAngeles" and groupid=2

docs/en/07-develop/03-insert-data/03-opentsdb-telnet.mdx

@@ -31,7 +31,7 @@ For example:
 ```txt
 meters.current 1648432611250 11.3 location=California.LosAngeles groupid=3
 ```
-
+- The defult child table name is generated by rules.You can configure smlChildTableName in taos.cfg to specify chile table names, for example, `smlChildTableName=tname`. You can insert `meters.current 1648432611250 11.3 tname=cpu1 location=California.LosAngeles groupid=3` and the cpu1 table will be automatically created. Note that if multiple rows have the same tname but different tag_set values, the tag_set of the first row is used to create the table and the others are ignored.
 Please refer to [OpenTSDB Telnet API](http://opentsdb.net/docs/build/html/api_telnet/put.html) for more details.
 
 ## Examples
@@ -79,3 +79,6 @@ taos> select tbname, * from `meters.current`;
  t_7e7b26dd860280242c6492a16... | 2022-03-28 09:56:51.250 |              12.600000000 | 2       | California.SanFrancisco               |
 Query OK, 4 row(s) in set (0.005399s)
 ```
+## Query Examples
+If you want query the data of `location=California.LosAngeles groupid=3`，here is the query sql：
+select * from `meters.voltage` where location="California.LosAngeles" and groupid=3

docs/en/07-develop/03-insert-data/04-opentsdb-json.mdx

@@ -48,7 +48,7 @@ Please refer to [OpenTSDB HTTP API](http://opentsdb.net/docs/build/html/api_http
 
 - In JSON protocol, strings will be converted to nchar type and numeric values will be converted to double type.
 - Only data in array format is accepted and so an array must be used even if there is only one row.
-
+- The defult child table name is generated by rules.You can configure smlChildTableName in taos.cfg to specify chile table names, for example, `smlChildTableName=tname`. You can insert `"tags": { "host": "web02","dc": "lga","tname":"cpu1"}` and the cpu1 table will be automatically created. Note that if multiple rows have the same tname but different tag_set values, the tag_set of the first row is used to create the table and the others are ignored.
 :::
 
 ## Examples
@@ -94,3 +94,6 @@ taos> select * from `meters.current`;
  2022-03-28 09:56:51.250 |              12.600000000 |               2.000000000 | California.SanFrancisco               |
 Query OK, 2 row(s) in set (0.004076s)
 ```
+## Query Examples
+If you want query the data of "tags": {"location": "California.LosAngeles", "groupid": 1}，here is the query sql：
+select * from `meters.voltage` where location="California.LosAngeles" and groupid=1

docs/en/12-taos-sql/06-select.md

@@ -181,6 +181,14 @@ In TDengine, the first column of all tables must be a timestamp. This column is
 select _rowts, max(current) from meters;
 ```
 
+**\_IROWTS**
+
+The \_IROWTS pseudocolumn can only be used with INTERP function. This pseudocolumn can be used to retrieve the corresponding timestamp column associated with the interpolation results.
+
+```sql
+select _irowts, interp(current) from meters range('2020-01-01 10:00:00', '2020-01-01 10:30:00') every(1s) fill(linear);
+```
+
 ## Query Objects
 
 `FROM` can be followed by a number of tables or super tables, or can be followed by a sub-query.

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

@@ -867,6 +867,7 @@ INTERP(expr)
 - The number of rows in the result set of `INTERP` is determined by the parameter `EVERY`. Starting from timestamp1, one interpolation is performed for every time interval specified `EVERY` parameter.
 - Interpolation is performed based on `FILL` parameter.
 - `INTERP` can only be used to interpolate in single timeline. So it must be used with `partition by tbname` when it's used on a STable.
+- Pseudo column `_irowts` can be used along with `INTERP` to return the timestamps associated with interpolation points(support after version 3.0.1.4).
 
 ### LAST
 

docs/en/12-taos-sql/20-keywords.md

@@ -343,6 +343,7 @@ The following list shows all reserved keywords:
 ### \_
 
 - \_C0
+- \_IROWTS
 - \_QDURATION
 - \_QEND
 - \_QSTART

docs/en/12-taos-sql/22-meta.md

@@ -11,7 +11,15 @@ TDengine includes a built-in database named `INFORMATION_SCHEMA` to provide acce
 4. Future versions of TDengine can add new columns to INFORMATION_SCHEMA tables without affecting existing business systems.
 5. It is easier for users coming from other database management systems. For example, Oracle users can query data dictionary tables.
 
-Note: SHOW statements are still supported for the convenience of existing users.
+:::info
+
+- SHOW statements are still supported for the convenience of existing users.
+- Some columns in the system table may be keywords, and you need to use the escape character '\`' when querying, for example, to query the VGROUPS in the database `test`:
+```sql 
+   select `vgroups` from ins_databases where name = 'test';
+``` 
+
+:::
 
 This document introduces the tables of INFORMATION_SCHEMA and their structure.
 
@@ -102,7 +110,11 @@ Provides information about user-created databases. Similar to SHOW DATABASES.
 | 24  | wal_retention_period | INT              | WAL retention period                                   |
 | 25  |  wal_retention_size  | INT              | Maximum WAL size                                   |
 | 26  |   wal_roll_period    | INT              | WAL rotation period                                 |
-| 27  |   wal_segment_size   | WAL file size |
+| 27  |   wal_segment_size   | BIGINT | WAL file size |
+| 28  |   stt_trigger   | SMALLINT | The threshold for number of files to trigger file merging |
+| 29  |   table_prefix   | SMALLINT | The prefix length in the table name that is ignored when distributing table to vnode based on table name |
+| 30  |   table_suffix   | SMALLINT | The suffix length in the table name that is ignored when distributing table to vnode based on table name |
+| 31  |   tsdb_pagesize   | INT | The page size for internal storage engine, its unit is KB |
 
 ## INS_FUNCTIONS
 

docs/en/12-taos-sql/29-changes.md

@@ -11,12 +11,13 @@ description: "This document explains how TDengine SQL has changed in version 3.0
 | 1 | VARCHAR | Added | Alias of BINARY.
 | 2 | TIMESTAMP literal | Added | TIMESTAMP 'timestamp format' syntax now supported.
 | 3 | _ROWTS pseudocolumn | Added | Indicates the primary key. Alias of _C0.
-| 4 | INFORMATION_SCHEMA | Added |	Database for system metadata containing all schema definitions
-| 5 | PERFORMANCE_SCHEMA | Added | Database for system performance information.
-| 6 | Connection queries | Deprecated | Connection queries are no longer supported. The syntax and interfaces are deprecated.
-| 7 | Mixed operations | Enhanced | Mixing scalar and vector operations in queries has been enhanced and is supported in all SELECT clauses.
-| 8 | Tag operations | Added | Tag columns can be used in queries and clauses like data columns.
-| 9 | Timeline clauses and time functions in supertables | Enhanced | When PARTITION BY is not used, data in supertables is merged into a single timeline.
+| 4 | _IROWTS pseudocolumn | Added | Used to retrieve timestamps with INTERP function.
+| 5 | INFORMATION_SCHEMA | Added |	Database for system metadata containing all schema definitions
+| 6 | PERFORMANCE_SCHEMA | Added | Database for system performance information.
+| 7 | Connection queries | Deprecated | Connection queries are no longer supported. The syntax and interfaces are deprecated.
+| 8 | Mixed operations | Enhanced | Mixing scalar and vector operations in queries has been enhanced and is supported in all SELECT clauses.
+| 9 | Tag operations | Added | Tag columns can be used in queries and clauses like data columns.
+| 10 | Timeline clauses and time functions in supertables | Enhanced | When PARTITION BY is not used, data in supertables is merged into a single timeline.
 
 ## SQL Syntax
 

docs/en/14-reference/12-config/index.md

@@ -177,12 +177,21 @@ The parameters described in this document by the effect that they have on the sy
 ### maxNumOfDistinctRes
 
 | Attribute     | Description                                                                                                                                                                         |
-| -------- | -------------------------------- | --- |
+| -------- | -------------------------------- |
 | Applicable    | Server Only                                                                                                                                                                         |
 | Meaning       | The maximum number of distinct rows returned |
 | Value Range   | [100,000 - 100,000,000]                      |
 | Default Value | 100,000                                      |
 
+### keepColumnName
+
+| Attribute     | Description                             |
+| -------- | -------------------------------- |
+| Applicable | Client only                     |
+| Meaning     | When the Last, First, LastRow function is queried, whether the returned column name contains the function name. |
+| Value Range | 0 means including the function name, 1 means not including the function name.     |
+| Default Value   | 0                            |
+
 ## Locale Parameters
 
 ### timezone

docs/en/14-reference/13-schemaless/13-schemaless.md

@@ -47,9 +47,8 @@ In the schemaless writing data line protocol, each data item in the field_set ne
 
 - `t`, `T`, `true`, `True`, `TRUE`, `f`, `F`, `false`, and `False` will be handled directly as BOOL types.
 
-For example, the following data rows indicate that the t1 label is "3" (NCHAR), the t2 label is "4" (NCHAR), and the t3 label
-is "t3" to the super table named `st` labeled "t3" (NCHAR), write c1 column as 3 (BIGINT), c2 column as false (BOOL), c3 column
-is "passit" (BINARY), c4 column is 4 (DOUBLE), and the primary key timestamp is 1626006833639000000 in one row.
+For example, the following data rows write c1 column as 3 (BIGINT), c2 column as false (BOOL), c3 column
+as "passit" (BINARY), c4 column as 4 (DOUBLE), and the primary key timestamp as 1626006833639000000 to child table with the t1 label as "3" (NCHAR), the t2 label as "4" (NCHAR), and the t3 label as "t3" (NCHAR) and the super table named `st`.
 
 ```json
 st,t1=3,t2=4,t3=t3 c1=3i64,c3="passit",c2=false,c4=4f64 1626006833639000000
@@ -69,7 +68,7 @@ Schemaless writes process row data according to the following principles.
 
 Note that tag_key1, tag_key2 are not the original order of the tags entered by the user but the result of using the tag names in ascending order of the strings. Therefore, tag_key1 is not the first tag entered in the line protocol.
 The string's MD5 hash value "md5_val" is calculated after the ranking is completed. The calculation result is then combined with the string to generate the table name: "t_md5_val". "t_" is a fixed prefix that every table generated by this mapping relationship has.
-You can configure smlChildTableName to specify table names, for example, `smlChildTableName=tname`. You can insert `st,tname=cpul,t1=4 c1=3 1626006833639000000` and the cpu1 table will be automatically created. Note that if multiple rows have the same tname but different tag_set values, the tag_set of the first row is used to create the table and the others are ignored.
+You can configure smlChildTableName in taos.cfg to specify table names, for example, `smlChildTableName=tname`. You can insert `st,tname=cpul,t1=4 c1=3 1626006833639000000` and the cpu1 table will be automatically created. Note that if multiple rows have the same tname but different tag_set values, the tag_set of the first row is used to create the table and the others are ignored.
 
 2. If the super table obtained by parsing the line protocol does not exist, this super table is created.
 3. If the subtable obtained by the parse line protocol does not exist, Schemaless creates the sub-table according to the subtable name determined in steps 1 or 2.
@@ -78,7 +77,7 @@ You can configure smlChildTableName to specify table names, for example, `smlChi
    NULL.
 6. For BINARY or NCHAR columns, if the length of the value provided in a data row exceeds the column type limit, the maximum length of characters allowed to be stored in the column is automatically increased (only incremented and not decremented) to ensure complete preservation of the data.
 7. Errors encountered throughout the processing will interrupt the writing process and return an error code.
-8. It is assumed that the order of field_set in a supertable is consistent, meaning that the first record contains all fields and subsequent records store fields in the same order. If the order is not consistent, set smlDataFormat to false. Otherwise, data will be written out of order and a database error will occur.
+8. It is assumed that the order of field_set in a supertable is consistent, meaning that the first record contains all fields and subsequent records store fields in the same order. If the order is not consistent, set smlDataFormat in taos.cfg to false. Otherwise, data will be written out of order and a database error will occur.(smlDataFormat in taos.cfg default to false after version of 3.0.1.3)
 
 :::tip
 All processing logic of schemaless will still follow TDengine's underlying restrictions on data structures, such as the total length of each row of data cannot exceed

docs/en/28-releases/01-tdengine.md

@@ -6,6 +6,10 @@ description: TDengine release history, Release Notes and download links.
 
 import Release from "/components/ReleaseV3";
 
+## 3.0.1.3
+
+<Release type="tdengine" version="3.0.1.3" />
+
 ## 3.0.1.2
 
 <Release type="tdengine" version="3.0.1.2" />

docs/en/28-releases/02-tools.md

@@ -6,6 +6,10 @@ description: taosTools release history, Release Notes, download links.
 
 import Release from "/components/ReleaseV3";
 
+## 2.2.3
+
+<Release type="tools" version="2.2.3" />
+
 ## 2.2.2
 
 <Release type="tools" version="2.2.2" />

docs/zh/07-develop/03-insert-data/02-influxdb-line.mdx

@@ -37,7 +37,9 @@ meters,location=California.LosAngeles,groupid=2 current=13.4,voltage=223,phase=0
 - tag_set 中的所有的数据自动转化为 nchar 数据类型;
 - field_set 中的每个数据项都需要对自身的数据类型进行描述, 比如 1.2f32 代表 float 类型的数值 1.2, 如果不带类型后缀会被当作 double 处理;
 - timestamp 支持多种时间精度。写入数据的时候需要用参数指定时间精度，支持从小时到纳秒的 6 种时间精度。
-
+- 为了提高写入的效率，默认假设同一个超级表中 field_set 的顺序是一样的（第一条数据包含所有的 field，后面的数据按照这个顺序），如果顺序不一样，需要配置参数 smlDataFormat 为 false，否则，数据写入按照相同顺序写入，库中数据会异常。（3.0.1.3之后的版本 smlDataFormat 默认为 false） [TDengine 无模式写入参考指南](/reference/schemaless/#无模式写入行协议)
+- 默认生产的子表名是根据规则生成的唯一ID值。为了让用户可以指定生成的表名，可以通过在taos.cfg里配置 smlChildTableName 参数来指定。
+举例如下：配置 smlChildTableName=tname 插入数据为 st,tname=cpu1,t1=4 c1=3 1626006833639000000 则创建的表名为 cpu1，注意如果多行数据 tname 相同，但是后面的 tag_set 不同，则使用第一行自动建表时指定的 tag_set，其他的行会忽略）。[TDengine 无模式写入参考指南](/reference/schemaless/#无模式写入行协议)
 :::
 
 要了解更多可参考：[InfluxDB Line 协议官方文档](https://docs.influxdata.com/influxdb/v2.0/reference/syntax/line-protocol/) 和 [TDengine 无模式写入参考指南](/reference/schemaless/#无模式写入行协议)
@@ -64,3 +66,7 @@ meters,location=California.LosAngeles,groupid=2 current=13.4,voltage=223,phase=0
     <CLine />
   </TabItem>
 </Tabs>
+
+## 查询示例
+比如查询 location=California.LosAngeles,groupid=2 子表的数据可以通过如下sql：
+select * from meters where location=California.LosAngeles and groupid=2

docs/zh/07-develop/03-insert-data/03-opentsdb-telnet.mdx

@@ -32,6 +32,8 @@ OpenTSDB 行协议同样采用一行字符串来表示一行数据。OpenTSDB
 meters.current 1648432611250 11.3 location=California.LosAngeles groupid=3
 ```
 
+- 默认生产的子表名是根据规则生成的唯一ID值。为了让用户可以指定生成的表名，可以通过在taos.cfg里配置 smlChildTableName 参数来指定。
+举例如下：配置 smlChildTableName=tname 插入数据为 meters.current 1648432611250 11.3 tname=cpu1 location=California.LosAngeles groupid=3 则创建的表名为 cpu1，注意如果多行数据 tname 相同，但是后面的 tag_set 不同，则使用第一行自动建表时指定的 tag_set，其他的行会忽略）。
 参考[OpenTSDB Telnet API 文档](http://opentsdb.net/docs/build/html/api_telnet/put.html)。
 
 ## 示例代码
@@ -79,3 +81,6 @@ taos> select tbname, * from `meters.current`;
  t_7e7b26dd860280242c6492a16... | 2022-03-28 09:56:51.250 |              12.600000000 | 2       | California.SanFrancisco               |
 Query OK, 4 row(s) in set (0.005399s)
 ```
+## 查询示例：
+想要查询 location=California.LosAngeles groupid=3 的数据，可以通过如下sql:
+select * from `meters.voltage` where location="California.LosAngeles" and groupid=3

docs/zh/07-develop/03-insert-data/04-opentsdb-json.mdx

@@ -48,7 +48,8 @@ OpenTSDB JSON 格式协议采用一个 JSON 字符串表示一行或多行数据
 
 - 对于 JSON 格式协议，TDengine 并不会自动把所有标签转成 nchar 类型， 字符串将将转为 nchar 类型， 数值将同样转换为 double 类型。
 - TDengine 只接收 JSON **数组格式**的字符串，即使一行数据也需要转换成数组形式。
-
+- 默认生产的子表名是根据规则生成的唯一ID值。为了让用户可以指定生成的表名，可以通过在taos.cfg里配置 smlChildTableName 参数来指定。
+举例如下：配置 smlChildTableName=tname 插入数据为 "tags": { "host": "web02","dc": "lga","tname":"cpu1"} 则创建的表名为 cpu1，注意如果多行数据 tname 相同，但是后面的 tag_set 不同，则使用第一行自动建表时指定的 tag_set，其他的行会忽略）。
 :::
 
 ## 示例代码
@@ -94,3 +95,7 @@ taos> select * from `meters.current`;
  2022-03-28 09:56:51.250 |              12.600000000 |               2.000000000 | California.SanFrancisco               |
 Query OK, 2 row(s) in set (0.004076s)
 ```
+
+## 查询示例
+想要查询"tags": {"location": "California.LosAngeles", "groupid": 1} 的数据，可以通过如下sql：
+select * from `meters.voltage` where location="California.LosAngeles" and groupid=1

docs/zh/12-taos-sql/06-select.md

@@ -182,6 +182,14 @@ TDengine 中，所有表的第一列都必须是时间戳类型，且为其主
 select _rowts, max(current) from meters;
 ```
 
+**\_IROWTS**
+
+\_irowts 伪列只能与 interp 函数一起使用，用于返回 interp 函数插值结果对应的时间戳列。
+
+```sql
+select _irowts, interp(current) from meters range('2020-01-01 10:00:00', '2020-01-01 10:30:00') every(1s) fill(linear);
+```
+
 ## 查询对象
 
 FROM 关键字后面可以是若干个表（超级表）列表，也可以是子查询的结果。

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

@@ -868,6 +868,7 @@ INTERP(expr)
 - INTERP 根据 EVERY 字段来确定输出时间范围内的结果条数，即从 timestamp1 开始每隔固定长度的时间（EVERY 值）进行插值。如果没有指定 EVERY，则默认窗口大小为无穷大，即从 timestamp1 开始只有一个窗口。
 - INTERP 根据 FILL 字段来决定在每个符合输出条件的时刻如何进行插值。
 - INTERP 只能在一个时间序列内进行插值，因此当作用于超级表时必须跟 partition by tbname 一起使用。
+- INTERP 可以与伪列 _irowts 一起使用，返回插值点所对应的时间戳(3.0.1.4版本以后支持)。
 
 ### LAST
 

docs/zh/12-taos-sql/20-keywords.md

@@ -344,6 +344,7 @@ description: TDengine 保留关键字的详细列表
 ### \_
 
 - \_C0
+- \_IROWTS
 - \_QDURATION
 - \_QEND
 - \_QSTART

docs/zh/12-taos-sql/22-meta.md

@@ -12,7 +12,15 @@ TDengine 内置了一个名为 `INFORMATION_SCHEMA` 的数据库，提供对数
 4. TDengine 在后续演进中可以灵活的添加已有 INFORMATION_SCHEMA 中表的列，而不用担心对既有业务系统造成影响
 5. 与其他数据库系统更具互操作性。例如，Oracle 数据库用户熟悉查询 Oracle 数据字典中的表
 
-Note: 由于 SHOW 语句已经被开发者熟悉和广泛使用，所以它们仍然被保留。
+:::info
+
+- 由于 SHOW 语句已经被开发者熟悉和广泛使用，所以它们仍然被保留。
+- 系统表中的一些列可能是关键字，在查询时需要使用转义符'\`'，例如查询数据库 test 有几个 VGROUP：
+```sql 
+   select `vgroups` from ins_databases where name = 'test';
+``` 
+
+:::
 
 本章将详细介绍 `INFORMATION_SCHEMA` 这个内置元数据库中的表和表结构。
 
@@ -103,7 +111,11 @@ Note: 由于 SHOW 语句已经被开发者熟悉和广泛使用，所以它们
 | 24  | wal_retention_period | INT              | WAL 的保存时长                                   |
 | 25  |  wal_retention_size  | INT              | WAL 的保存上限                                   |
 | 26  |   wal_roll_period    | INT              | wal 文件切换时长                                 |
-| 27  |   wal_segment_size   | wal 单个文件大小 |
+| 27  |   wal_segment_size   | BIGINT | wal 单个文件大小 |
+| 28  |   stt_trigger   | SMALLINT | 触发文件合并的落盘文件的个数 |
+| 29  |   table_prefix   | SMALLINT | 内部存储引擎根据表名分配存储该表数据的 VNODE 时要忽略的前缀的长度 |
+| 30  |   table_suffix   | SMALLINT | 内部存储引擎根据表名分配存储该表数据的 VNODE 时要忽略的后缀的长度 |
+| 31  |   tsdb_pagesize   | INT | 时序数据存储引擎中的页大小 |
 
 ## INS_FUNCTIONS
 

docs/zh/12-taos-sql/29-changes.md

@@ -11,12 +11,13 @@ description: "TDengine 3.0 版本的语法变更说明"
 | 1 | VARCHAR | 新增 | BINARY类型的别名。
 | 2 | TIMESTAMP字面量 | 新增 | 新增支持 TIMESTAMP 'timestamp format' 语法。
 | 3 | _ROWTS伪列 | 新增 | 表示时间戳主键。是_C0伪列的别名。
-| 4 | INFORMATION_SCHEMA | 新增 |	包含各种SCHEMA定义的系统数据库。
-| 5 | PERFORMANCE_SCHEMA | 新增 | 包含运行信息的系统数据库。
-| 6 | 连续查询 | 废除 | 不再支持连续查询。相关的各种语法和接口废除。
-| 7 | 混合运算 | 增强 | 查询中的混合运算（标量运算和矢量运算混合）全面增强，SELECT的各个子句均全面支持符合语法语义的混合运算。
-| 8 | 标签运算 | 新增 |在查询中，标签列可以像普通列一样参与各种运算，用于各种子句。
-| 9 | 时间线子句和时间函数用于超级表查询 | 增强 |没有PARTITION BY时，超级表的数据会被合并成一条时间线。
+| 4 | _IROWTS伪列 | 新增 | 用于返回 interp 函数插值结果对应的时间戳列。
+| 5 | INFORMATION_SCHEMA | 新增 |	包含各种SCHEMA定义的系统数据库。
+| 6 | PERFORMANCE_SCHEMA | 新增 | 包含运行信息的系统数据库。
+| 7 | 连续查询 | 废除 | 不再支持连续查询。相关的各种语法和接口废除。
+| 8 | 混合运算 | 增强 | 查询中的混合运算（标量运算和矢量运算混合）全面增强，SELECT的各个子句均全面支持符合语法语义的混合运算。
+| 9 | 标签运算 | 新增 |在查询中，标签列可以像普通列一样参与各种运算，用于各种子句。
+| 10 | 时间线子句和时间函数用于超级表查询 | 增强 |没有PARTITION BY时，超级表的数据会被合并成一条时间线。
 
 ## SQL 语句变更
 

docs/zh/14-reference/12-config/index.md

@@ -177,12 +177,21 @@ taos --dump-config
 ### maxNumOfDistinctRes
 
 | 属性     | 说明                             |
-| -------- | -------------------------------- | --- |
+| -------- | -------------------------------- |
 | 适用范围 | 仅服务端适用                     |
 | 含义     | 允许返回的 distinct 结果最大行数 |
 | 取值范围 | 默认值为 10 万，最大值 1 亿      |
 | 缺省值   | 10 万                            |
 
+### keepColumnName
+
+| 属性     | 说明                             |
+| -------- | -------------------------------- |
+| 适用范围 | 仅客户端适用                     |
+| 含义     | Last、First、LastRow 函数查询时，返回的列名是否包含函数名。 |
+| 取值范围 | 0 表示包含函数名，1 表示不包含函数名。      |
+| 缺省值   | 0                            |
+
 ## 区域相关
 
 ### timezone

docs/zh/14-reference/13-schemaless/13-schemaless.md

@@ -67,9 +67,12 @@ st,t1=3,t2=4,t3=t3 c1=3i64,c3="passit",c2=false,c4=4f64 1626006833639000000
 "measurement,tag_key1=tag_value1,tag_key2=tag_value2"
 ```
 
+:::tip
 需要注意的是，这里的 tag_key1, tag_key2 并不是用户输入的标签的原始顺序，而是使用了标签名称按照字符串升序排列后的结果。所以，tag_key1 并不是在行协议中输入的第一个标签。
 排列完成以后计算该字符串的 MD5 散列值 "md5_val"。然后将计算的结果与字符串组合生成表名：“t_md5_val”。其中的 “t_” 是固定的前缀，每个通过该映射关系自动生成的表都具有该前缀。
-为了让用户可以指定生成的表名，可以通过配置 smlChildTableName 来指定（比如 配置 smlChildTableName=tname 插入数据为 st,tname=cpu1,t1=4 c1=3 1626006833639000000 则创建的表名为 cpu1，注意如果多行数据 tname 相同，但是后面的 tag_set 不同，则使用第一次自动建表时指定的 tag_set，其他的会忽略）。
+:::tip 
+为了让用户可以指定生成的表名，可以通过在taos.cfg里配置 smlChildTableName 参数来指定。
+举例如下：配置 smlChildTableName=tname 插入数据为 st,tname=cpu1,t1=4 c1=3 1626006833639000000 则创建的表名为 cpu1，注意如果多行数据 tname 相同，但是后面的 tag_set 不同，则使用第一行自动建表时指定的 tag_set，其他的行会忽略）。
 
 2. 如果解析行协议获得的超级表不存在，则会创建这个超级表（不建议手动创建超级表，不然插入数据可能异常）。
 3. 如果解析行协议获得子表不存在，则 Schemaless 会按照步骤 1 或 2 确定的子表名来创建子表。

docs/zh/28-releases/01-tdengine.md

@@ -6,6 +6,10 @@ description: TDengine 发布历史、Release Notes 及下载链接
 
 import Release from "/components/ReleaseV3";
 
+## 3.0.1.3
+
+<Release type="tdengine" version="3.0.1.3" />
+
 ## 3.0.1.2
 
 <Release type="tdengine" version="3.0.1.2" />

docs/zh/28-releases/02-tools.md

@@ -6,6 +6,10 @@ description: taosTools 的发布历史、Release Notes 和下载链接
 
 import Release from "/components/ReleaseV3";
 
+## 2.2.3
+
+<Release type="tools" version="2.2.3" />
+
 ## 2.2.2
 
 <Release type="tools" version="2.2.2" />

examples/lua/README.md

 # TDengine driver connector for Lua
 
 It's a Lua implementation for [TDengine](https://github.com/taosdata/TDengine), an open-sourced big data platform designed and optimized for the Internet of Things (IoT), Connected Cars, Industrial IoT, and IT Infrastructure and Application Monitoring. You may need to install Lua5.3 .
+As TDengine is built with lua-enable with default configure, the built-in lua lib conflicts with external lua lib. The following commands require TDengine built with lua-disable.
+To disable built-in lua:
+```
+mkdir debug && cd debug
+cmake .. -DBUILD_LUA=false && cmake --build .
+```
 
 ## Lua Dependencies
 - Lua: 

examples/lua/lua51/lua_connector51.c

@@ -29,7 +29,7 @@ static int l_connect(lua_State *L){
   luaL_checktype(L, 1, LUA_TTABLE);
 
   lua_getfield(L, 1,"host");
-  if (lua_isstring(L,-1)){
+  if (lua_isstring(L, -1)){
     host = lua_tostring(L, -1);
     // printf("host = %s\n", host);
   }
@@ -58,8 +58,9 @@ static int l_connect(lua_State *L){
     //printf("password = %s\n", password);
   }
 
-  lua_settop(L,0);
+  lua_settop(L, 0);
 
+  taos_init();
   
   lua_newtable(L);
   int table_index = lua_gettop(L);
@@ -125,7 +126,7 @@ static int l_query(lua_State *L){
       //printf("row index:%d\n",rows);
       rows++;
 
-      lua_pushnumber(L,rows);
+      lua_pushnumber(L, rows);
       lua_newtable(L);
 
       for (int i = 0; i < num_fields; ++i) {
@@ -136,15 +137,19 @@ static int l_query(lua_State *L){
 	lua_pushstring(L,fields[i].name);
 	int32_t* length = taos_fetch_lengths(result);
 	switch (fields[i].type) {
+	case TSDB_DATA_TYPE_UTINYINT:
 	case TSDB_DATA_TYPE_TINYINT:
 	  lua_pushinteger(L,*((char *)row[i]));
 	  break;
+	case TSDB_DATA_TYPE_USMALLINT:
 	case TSDB_DATA_TYPE_SMALLINT:
 	  lua_pushinteger(L,*((short *)row[i]));
 	  break;
+	case TSDB_DATA_TYPE_UINT:
 	case TSDB_DATA_TYPE_INT:
 	  lua_pushinteger(L,*((int *)row[i]));
 	  break;
+	case TSDB_DATA_TYPE_UBIGINT:
 	case TSDB_DATA_TYPE_BIGINT:
 	  lua_pushinteger(L,*((int64_t *)row[i]));
 	  break;
@@ -154,6 +159,7 @@ static int l_query(lua_State *L){
 	case TSDB_DATA_TYPE_DOUBLE:
 	  lua_pushnumber(L,*((double *)row[i]));
 	  break;
+	case TSDB_DATA_TYPE_JSON:
 	case TSDB_DATA_TYPE_BINARY:
 	case TSDB_DATA_TYPE_NCHAR:
 	  //printf("type:%d, max len:%d, current len:%d\n",fields[i].type, fields[i].bytes, length[i]);
@@ -197,7 +203,7 @@ void async_query_callback(void *param, TAOS_RES *result, int code){
     printf("failed, reason:%s\n", taos_errstr(result));
     lua_pushinteger(L, -1);
     lua_setfield(L, table_index, "code");    
-    lua_pushstring(L,"something is wrong");// taos_errstr(taos));
+    lua_pushstring(L, taos_errstr(result));
     lua_setfield(L, table_index, "error");    
   }else{
     //printf("success to async query.\n");
@@ -214,9 +220,9 @@ void async_query_callback(void *param, TAOS_RES *result, int code){
 
 static int l_async_query(lua_State *L){
   int r = luaL_ref(L, LUA_REGISTRYINDEX);
-  TAOS *    taos = (TAOS*)lua_topointer(L,1);
-  const char * sqlstr = lua_tostring(L,2);
-  // int stime = luaL_checknumber(L,3);
+  TAOS *    taos = (TAOS*)lua_topointer(L, 1);
+  const char * sqlstr = lua_tostring(L, 2);
+  // int stime = luaL_checknumber(L, 3);
 
   lua_newtable(L);
   int table_index = lua_gettop(L);
@@ -224,7 +230,7 @@ static int l_async_query(lua_State *L){
   struct async_query_callback_param *p = malloc(sizeof(struct async_query_callback_param));
   p->state = L;
   p->callback=r;
-  // printf("r:%d, L:%d\n",r,L);
+  // printf("r:%d, L:%d\n", r, L);
   taos_query_a(taos,sqlstr,async_query_callback,p);
 
   lua_pushnumber(L, 0);
@@ -267,7 +273,7 @@ static const struct luaL_Reg lib[] = {
 extern int luaopen_luaconnector51(lua_State* L)
 {
   //  luaL_register(L, "luaconnector51", lib);
-  lua_newtable (L);
+  lua_newtable(L);
   luaL_setfuncs(L,lib,0);
   return 1;
 }

examples/lua/lua_connector.c

@@ -29,7 +29,7 @@ static int l_connect(lua_State *L){
   luaL_checktype(L, 1, LUA_TTABLE);
 
   lua_getfield(L, 1,"host");
-  if (lua_isstring(L,-1)){
+  if (lua_isstring(L, -1)){
     host = lua_tostring(L, -1);
     // printf("host = %s\n", host);
   }
@@ -58,7 +58,7 @@ static int l_connect(lua_State *L){
     //printf("password = %s\n", password);
   }
 
-  lua_settop(L,0);
+  lua_settop(L, 0);
 
   taos_init();
   
@@ -126,7 +126,7 @@ static int l_query(lua_State *L){
       //printf("row index:%d\n",rows);
       rows++;
 
-      lua_pushnumber(L,rows);
+      lua_pushnumber(L, rows);
       lua_newtable(L);
 
       for (int i = 0; i < num_fields; ++i) {
@@ -203,7 +203,7 @@ void async_query_callback(void *param, TAOS_RES *result, int code){
     printf("failed, reason:%s\n", taos_errstr(result));
     lua_pushinteger(L, -1);
     lua_setfield(L, table_index, "code");    
-    lua_pushstring(L,"something is wrong");// taos_errstr(taos));
+    lua_pushstring(L, taos_errstr(result));
     lua_setfield(L, table_index, "error");    
   }else{
     //printf("success to async query.\n");
@@ -220,9 +220,9 @@ void async_query_callback(void *param, TAOS_RES *result, int code){
 
 static int l_async_query(lua_State *L){
   int r = luaL_ref(L, LUA_REGISTRYINDEX);
-  TAOS *    taos = (TAOS*)lua_topointer(L,1);
-  const char * sqlstr = lua_tostring(L,2);
-  // int stime = luaL_checknumber(L,3);
+  TAOS *    taos = (TAOS*)lua_topointer(L, 1);
+  const char * sqlstr = lua_tostring(L, 2);
+  // int stime = luaL_checknumber(L, 3);
 
   lua_newtable(L);
   int table_index = lua_gettop(L);
@@ -230,7 +230,7 @@ static int l_async_query(lua_State *L){
   struct async_query_callback_param *p = malloc(sizeof(struct async_query_callback_param));
   p->state = L;
   p->callback=r;
-  // printf("r:%d, L:%d\n",r,L);
+  // printf("r:%d, L:%d\n", r, L);
   taos_query_a(taos,sqlstr,async_query_callback,p);
 
   lua_pushnumber(L, 0);

examples/lua/test.lua

@@ -176,8 +176,14 @@ end
 driver.query_a(conn,"INSERT INTO therm1 VALUES ('2019-09-01 00:00:00.005', 100),('2019-09-01 00:00:00.006', 101),('2019-09-01 00:00:00.007', 102)", async_query_callback)
 
 res = driver.query(conn, "create stream stream_avg_degree into avg_degree as select avg(degree) from thermometer interval(5s) sliding(1s)")
+if res.code ~=0 then
+   print("create stream--- failed:"..res.error)
+   return
+else
+   print("create stream--- pass")
+end
 
-print("From now on we start continous insert in an definite loop, pls wait for about 10 seconds and check stream table for result.")
+print("From now on we start continous insertion in an definite loop, please wait for about 10 seconds and check stream table avg_degree for result.")
 local loop_index = 0
 while loop_index < 10 do
    local t = os.time()*1000
@@ -193,5 +199,5 @@ while loop_index < 10 do
    os.execute("sleep " .. 1)
    loop_index = loop_index + 1
 end
-driver.query(conn,"DROP STREAM IF EXISTS avg_therm_s")
+driver.query(conn,"DROP STREAM IF EXISTS stream_avg_degree")
 driver.close(conn)

include/common/tglobal.h

@@ -98,6 +98,7 @@ extern int32_t tsQueryRsmaTolerance;
 extern bool    tsQueryPlannerTrace;
 extern int32_t tsQueryNodeChunkSize;
 extern bool    tsQueryUseNodeAllocator;
+extern bool    tsKeepColumnName;
 
 // client
 extern int32_t tsMinSlidingTime;

include/common/tmsg.h

@@ -55,11 +55,10 @@ extern int32_t tMsgDict[];
 
 #define TMSG_SEG_CODE(TYPE) (((TYPE)&0xff00) >> 8)
 #define TMSG_SEG_SEQ(TYPE)  ((TYPE)&0xff)
-#define TMSG_INFO(TYPE)                                                                                      \
-  ((TYPE) >= 0 && ((TYPE) < TDMT_DND_MAX_MSG || (TYPE) < TDMT_MND_MAX_MSG || (TYPE) < TDMT_VND_MAX_MSG ||    \
-                   (TYPE) < TDMT_SCH_MAX_MSG || (TYPE) < TDMT_STREAM_MAX_MSG || (TYPE) < TDMT_MON_MAX_MSG || \
-                   (TYPE) < TDMT_SYNC_MAX_MSG))                                                              \
-      ? tMsgInfo[tMsgDict[TMSG_SEG_CODE(TYPE)] + TMSG_SEG_SEQ(TYPE)]                                         \
+#define TMSG_INFO(TYPE)                                                                                                \
+  ((TYPE) < TDMT_DND_MAX_MSG || (TYPE) < TDMT_MND_MAX_MSG || (TYPE) < TDMT_VND_MAX_MSG || (TYPE) < TDMT_SCH_MAX_MSG || \
+   (TYPE) < TDMT_STREAM_MAX_MSG || (TYPE) < TDMT_MON_MAX_MSG || (TYPE) < TDMT_SYNC_MAX_MSG)                            \
+      ? tMsgInfo[tMsgDict[TMSG_SEG_CODE(TYPE)] + TMSG_SEG_SEQ(TYPE)]                                                   \
       : 0
 #define TMSG_INDEX(TYPE) (tMsgDict[TMSG_SEG_CODE(TYPE)] + TMSG_SEG_SEQ(TYPE))
 

include/common/ttokendef.h

@@ -226,110 +226,111 @@
 #define TK_WSTART               208
 #define TK_WEND                 209
 #define TK_WDURATION            210
-#define TK_CAST                 211
-#define TK_NOW                  212
-#define TK_TODAY                213
-#define TK_TIMEZONE             214
-#define TK_CLIENT_VERSION       215
-#define TK_SERVER_VERSION       216
-#define TK_SERVER_STATUS        217
-#define TK_CURRENT_USER         218
-#define TK_COUNT                219
-#define TK_LAST_ROW             220
-#define TK_CASE                 221
-#define TK_END                  222
-#define TK_WHEN                 223
-#define TK_THEN                 224
-#define TK_ELSE                 225
-#define TK_BETWEEN              226
-#define TK_IS                   227
-#define TK_NK_LT                228
-#define TK_NK_GT                229
-#define TK_NK_LE                230
-#define TK_NK_GE                231
-#define TK_NK_NE                232
-#define TK_MATCH                233
-#define TK_NMATCH               234
-#define TK_CONTAINS             235
-#define TK_IN                   236
-#define TK_JOIN                 237
-#define TK_INNER                238
-#define TK_SELECT               239
-#define TK_DISTINCT             240
-#define TK_WHERE                241
-#define TK_PARTITION            242
-#define TK_BY                   243
-#define TK_SESSION              244
-#define TK_STATE_WINDOW         245
-#define TK_SLIDING              246
-#define TK_FILL                 247
-#define TK_VALUE                248
-#define TK_NONE                 249
-#define TK_PREV                 250
-#define TK_LINEAR               251
-#define TK_NEXT                 252
-#define TK_HAVING               253
-#define TK_RANGE                254
-#define TK_EVERY                255
-#define TK_ORDER                256
-#define TK_SLIMIT               257
-#define TK_SOFFSET              258
-#define TK_LIMIT                259
-#define TK_OFFSET               260
-#define TK_ASC                  261
-#define TK_NULLS                262
-#define TK_ABORT                263
-#define TK_AFTER                264
-#define TK_ATTACH               265
-#define TK_BEFORE               266
-#define TK_BEGIN                267
-#define TK_BITAND               268
-#define TK_BITNOT               269
-#define TK_BITOR                270
-#define TK_BLOCKS               271
-#define TK_CHANGE               272
-#define TK_COMMA                273
-#define TK_COMPACT              274
-#define TK_CONCAT               275
-#define TK_CONFLICT             276
-#define TK_COPY                 277
-#define TK_DEFERRED             278
-#define TK_DELIMITERS           279
-#define TK_DETACH               280
-#define TK_DIVIDE               281
-#define TK_DOT                  282
-#define TK_EACH                 283
-#define TK_FAIL                 284
-#define TK_FILE                 285
-#define TK_FOR                  286
-#define TK_GLOB                 287
-#define TK_ID                   288
-#define TK_IMMEDIATE            289
-#define TK_IMPORT               290
-#define TK_INITIALLY            291
-#define TK_INSTEAD              292
-#define TK_ISNULL               293
-#define TK_KEY                  294
-#define TK_NK_BITNOT            295
-#define TK_NK_SEMI              296
-#define TK_NOTNULL              297
-#define TK_OF                   298
-#define TK_PLUS                 299
-#define TK_PRIVILEGE            300
-#define TK_RAISE                301
-#define TK_REPLACE              302
-#define TK_RESTRICT             303
-#define TK_ROW                  304
-#define TK_SEMI                 305
-#define TK_STAR                 306
-#define TK_STATEMENT            307
-#define TK_STRING               308
-#define TK_TIMES                309
-#define TK_UPDATE               310
-#define TK_VALUES               311
-#define TK_VARIABLE             312
-#define TK_VIEW                 313
-#define TK_WAL                  314
+#define TK_IROWTS               211
+#define TK_CAST                 212
+#define TK_NOW                  213
+#define TK_TODAY                214
+#define TK_TIMEZONE             215
+#define TK_CLIENT_VERSION       216
+#define TK_SERVER_VERSION       217
+#define TK_SERVER_STATUS        218
+#define TK_CURRENT_USER         219
+#define TK_COUNT                220
+#define TK_LAST_ROW             221
+#define TK_CASE                 222
+#define TK_END                  223
+#define TK_WHEN                 224
+#define TK_THEN                 225
+#define TK_ELSE                 226
+#define TK_BETWEEN              227
+#define TK_IS                   228
+#define TK_NK_LT                229
+#define TK_NK_GT                230
+#define TK_NK_LE                231
+#define TK_NK_GE                232
+#define TK_NK_NE                233
+#define TK_MATCH                234
+#define TK_NMATCH               235
+#define TK_CONTAINS             236
+#define TK_IN                   237
+#define TK_JOIN                 238
+#define TK_INNER                239
+#define TK_SELECT               240
+#define TK_DISTINCT             241
+#define TK_WHERE                242
+#define TK_PARTITION            243
+#define TK_BY                   244
+#define TK_SESSION              245
+#define TK_STATE_WINDOW         246
+#define TK_SLIDING              247
+#define TK_FILL                 248
+#define TK_VALUE                249
+#define TK_NONE                 250
+#define TK_PREV                 251
+#define TK_LINEAR               252
+#define TK_NEXT                 253
+#define TK_HAVING               254
+#define TK_RANGE                255
+#define TK_EVERY                256
+#define TK_ORDER                257
+#define TK_SLIMIT               258
+#define TK_SOFFSET              259
+#define TK_LIMIT                260
+#define TK_OFFSET               261
+#define TK_ASC                  262
+#define TK_NULLS                263
+#define TK_ABORT                264
+#define TK_AFTER                265
+#define TK_ATTACH               266
+#define TK_BEFORE               267
+#define TK_BEGIN                268
+#define TK_BITAND               269
+#define TK_BITNOT               270
+#define TK_BITOR                271
+#define TK_BLOCKS               272
+#define TK_CHANGE               273
+#define TK_COMMA                274
+#define TK_COMPACT              275
+#define TK_CONCAT               276
+#define TK_CONFLICT             277
+#define TK_COPY                 278
+#define TK_DEFERRED             279
+#define TK_DELIMITERS           280
+#define TK_DETACH               281
+#define TK_DIVIDE               282
+#define TK_DOT                  283
+#define TK_EACH                 284
+#define TK_FAIL                 285
+#define TK_FILE                 286
+#define TK_FOR                  287
+#define TK_GLOB                 288
+#define TK_ID                   289
+#define TK_IMMEDIATE            290
+#define TK_IMPORT               291
+#define TK_INITIALLY            292
+#define TK_INSTEAD              293
+#define TK_ISNULL               294
+#define TK_KEY                  295
+#define TK_NK_BITNOT            296
+#define TK_NK_SEMI              297
+#define TK_NOTNULL              298
+#define TK_OF                   299
+#define TK_PLUS                 300
+#define TK_PRIVILEGE            301
+#define TK_RAISE                302
+#define TK_REPLACE              303
+#define TK_RESTRICT             304
+#define TK_ROW                  305
+#define TK_SEMI                 306
+#define TK_STAR                 307
+#define TK_STATEMENT            308
+#define TK_STRING               309
+#define TK_TIMES                310
+#define TK_UPDATE               311
+#define TK_VALUES               312
+#define TK_VARIABLE             313
+#define TK_VIEW                 314
+#define TK_WAL                  315
 
 #define TK_NK_SPACE   300
 #define TK_NK_COMMENT 301

include/libs/executor/executor.h

@@ -136,7 +136,7 @@ int32_t qGetQueryTableSchemaVersion(qTaskInfo_t tinfo, char* dbName, char* table
  * @param handle
  * @return
  */
-int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, SLocalFetch *pLocal);
+int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, bool* hasMore, SLocalFetch *pLocal);
 int32_t qExecTask(qTaskInfo_t tinfo, SSDataBlock** pBlock, uint64_t* useconds);
 
 /**

include/libs/function/functionMgt.h

@@ -119,9 +119,10 @@ typedef enum EFunctionType {
   FUNCTION_TYPE_WSTART,
   FUNCTION_TYPE_WEND,
   FUNCTION_TYPE_WDURATION,
+  FUNCTION_TYPE_IROWTS,
 
   // internal function
-  FUNCTION_TYPE_SELECT_VALUE,
+  FUNCTION_TYPE_SELECT_VALUE = 3750,
   FUNCTION_TYPE_BLOCK_DIST,       // block distribution aggregate function
   FUNCTION_TYPE_BLOCK_DIST_INFO,  // block distribution pseudo column function
   FUNCTION_TYPE_TO_COLUMN,
@@ -212,6 +213,7 @@ bool fmIsClientPseudoColumnFunc(int32_t funcId);
 bool fmIsMultiRowsFunc(int32_t funcId);
 bool fmIsKeepOrderFunc(int32_t funcId);
 bool fmIsCumulativeFunc(int32_t funcId);
+bool fmIsInterpPseudoColumnFunc(int32_t funcId);
 
 int32_t fmGetDistMethod(const SFunctionNode* pFunc, SFunctionNode** pPartialFunc, SFunctionNode** pMergeFunc);
 

include/libs/nodes/nodes.h

@@ -239,6 +239,7 @@ typedef enum ENodeType {
   QUERY_NODE_PHYSICAL_PLAN_STREAM_FINAL_INTERVAL,
   QUERY_NODE_PHYSICAL_PLAN_STREAM_SEMI_INTERVAL,
   QUERY_NODE_PHYSICAL_PLAN_FILL,
+  QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL,
   QUERY_NODE_PHYSICAL_PLAN_MERGE_SESSION,
   QUERY_NODE_PHYSICAL_PLAN_STREAM_SESSION,
   QUERY_NODE_PHYSICAL_PLAN_STREAM_SEMI_SESSION,

include/libs/nodes/plannodes.h

@@ -464,6 +464,8 @@ typedef struct SFillPhysiNode {
   EOrder      inputTsOrder;
 } SFillPhysiNode;
 
+typedef SFillPhysiNode SStreamFillPhysiNode;
+
 typedef struct SMultiTableIntervalPhysiNode {
   SIntervalPhysiNode interval;
   SNodeList*         pPartitionKeys;

include/libs/stream/streamState.h

@@ -31,6 +31,7 @@ typedef struct {
   TDB*         db;
   TTB*         pStateDb;
   TTB*         pFuncStateDb;
+  TTB*         pFillStateDb;  // todo refactor
   TXN          txn;
 } SStreamState;
 
@@ -51,15 +52,22 @@ int32_t streamStateFuncDel(SStreamState* pState, const STupleKey* key);
 int32_t streamStatePut(SStreamState* pState, const SWinKey* key, const void* value, int32_t vLen);
 int32_t streamStateGet(SStreamState* pState, const SWinKey* key, void** pVal, int32_t* pVLen);
 int32_t streamStateDel(SStreamState* pState, const SWinKey* key);
+
+int32_t streamStateFillPut(SStreamState* pState, const SWinKey* key, const void* value, int32_t vLen);
+int32_t streamStateFillGet(SStreamState* pState, const SWinKey* key, void** pVal, int32_t* pVLen);
+int32_t streamStateFillDel(SStreamState* pState, const SWinKey* key);
+
 int32_t streamStateAddIfNotExist(SStreamState* pState, const SWinKey* key, void** pVal, int32_t* pVLen);
 int32_t streamStateReleaseBuf(SStreamState* pState, const SWinKey* key, void* pVal);
 void    streamFreeVal(void* val);
 
 SStreamStateCur* streamStateGetCur(SStreamState* pState, const SWinKey* key);
-SStreamStateCur* streamStateSeekKeyNext(SStreamState* pState, const SWinKey* key);
-SStreamStateCur* streamStateSeekKeyPrev(SStreamState* pState, const SWinKey* key);
+SStreamStateCur* streamStateGetAndCheckCur(SStreamState* pState, SWinKey* key);
+SStreamStateCur* streamStateFillSeekKeyNext(SStreamState* pState, const SWinKey* key);
+SStreamStateCur* streamStateFillSeekKeyPrev(SStreamState* pState, const SWinKey* key);
 void             streamStateFreeCur(SStreamStateCur* pCur);
 
+int32_t streamStateGetGroupKVByCur(SStreamStateCur* pCur, SWinKey* pKey, const void** pVal, int32_t* pVLen);
 int32_t streamStateGetKVByCur(SStreamStateCur* pCur, SWinKey* pKey, const void** pVal, int32_t* pVLen);
 
 int32_t streamStateSeekFirst(SStreamState* pState, SStreamStateCur* pCur);

packaging/cfg/nginxd.service

-[Unit]
-Description=Nginx For TDengine Service
-After=network-online.target
-Wants=network-online.target
-
-[Service]
-Type=forking
-PIDFile=/usr/local/nginxd/logs/nginx.pid
-ExecStart=/usr/local/nginxd/sbin/nginx
-ExecStop=/usr/local/nginxd/sbin/nginx -s stop
-TimeoutStopSec=1000000s
-LimitNOFILE=infinity
-LimitNPROC=infinity
-LimitCORE=infinity
-TimeoutStartSec=0
-StandardOutput=null
-Restart=always
-StartLimitBurst=3
-StartLimitInterval=60s
-
-[Install]
-WantedBy=multi-user.target

packaging/check_package.sh

@@ -38,8 +38,6 @@ temp_version=""
 fin_result=""
 
 service_config_dir="/etc/systemd/system"
-nginx_port=6060
-nginx_dir="/usr/local/nginxd"
 
 # Color setting
 RED='\033[0;31m'
@@ -132,10 +130,7 @@ function check_main_path() {
         check_file ${install_main_dir} $i
     done
     if [ "$verMode" == "cluster" ]; then
-        nginx_main_dir=("admin" "conf" "html" "sbin" "logs")
-        for i in "${nginx_main_dir[@]}";do
-            check_file ${nginx_dir}  $i
-        done
+        check_file ${install_main_dir} "share/admin"
     fi
     echo -e "Check main path:\033[32mOK\033[0m!"
 }
@@ -150,9 +145,6 @@ function check_bin_path() {
     for i in "${lbin_dir[@]}";do
         check_link ${bin_link_dir}/$i
     done
-    if [ "$verMode" == "cluster" ]; then
-        check_file ${nginx_dir}/sbin nginx
-    fi
     echo -e "Check bin  path:\033[32mOK\033[0m!"
 }
 

packaging/tools/install.sh

@@ -50,8 +50,7 @@ install_main_dir=${installDir}
 bin_dir="${installDir}/bin"
 
 service_config_dir="/etc/systemd/system"
-nginx_port=6060
-nginx_dir="/usr/local/nginxd"
+web_port=6041
 
 # Color setting
 RED='\033[0;31m'
@@ -182,7 +181,7 @@ function install_main_path() {
   ${csudo}mkdir -p ${install_main_dir}/include
   #  ${csudo}mkdir -p ${install_main_dir}/init.d
   if [ "$verMode" == "cluster" ]; then
-    ${csudo}mkdir -p ${nginx_dir}
+    ${csudo}mkdir -p ${install_main_dir}/share
   fi
 
   if [[ -e ${script_dir}/email ]]; then
@@ -218,12 +217,6 @@ function install_bin() {
   [ -x ${install_main_dir}/bin/TDinsight.sh ] && ${csudo}ln -s ${install_main_dir}/bin/TDinsight.sh ${bin_link_dir}/TDinsight.sh || :
   [ -x ${install_main_dir}/bin/remove.sh ] && ${csudo}ln -s ${install_main_dir}/bin/remove.sh ${bin_link_dir}/${uninstallScript} || :
   [ -x ${install_main_dir}/bin/set_core.sh ] && ${csudo}ln -s ${install_main_dir}/bin/set_core.sh ${bin_link_dir}/set_core || :
-
-  if [ "$verMode" == "cluster" ]; then
-    ${csudo}cp -r ${script_dir}/nginxd/* ${nginx_dir} && ${csudo}chmod 0555 ${nginx_dir}/*
-    ${csudo}mkdir -p ${nginx_dir}/logs
-    ${csudo}chmod 777 ${nginx_dir}/sbin/nginx
-  fi
 }
 
 function install_lib() {
@@ -574,6 +567,13 @@ function install_examples() {
   fi
 }
 
+function install_web() {
+  if [ -d "${script_dir}/share" ]; then
+    ${csudo}cp -rf ${script_dir}/share/* ${install_main_dir}/share
+  fi
+}
+
+
 function clean_service_on_sysvinit() {
   if pidof ${serverName} &>/dev/null; then
     ${csudo}service ${serverName} stop || :
@@ -654,16 +654,6 @@ function clean_service_on_systemd() {
   fi
   ${csudo}systemctl disable tarbitratord &>/dev/null || echo &>/dev/null
   ${csudo}rm -f ${tarbitratord_service_config}
-
-  if [ "$verMode" == "cluster" ]; then
-    nginx_service_config="${service_config_dir}/nginxd.service"
-    if systemctl is-active --quiet nginxd; then
-      echo "Nginx for ${productName} is running, stopping it..."
-      ${csudo}systemctl stop nginxd &>/dev/null || echo &>/dev/null
-    fi
-    ${csudo}systemctl disable nginxd &>/dev/null || echo &>/dev/null
-    ${csudo}rm -f ${nginx_service_config}
-  fi
 }
 
 function install_service_on_systemd() {
@@ -677,19 +667,6 @@ function install_service_on_systemd() {
   ${csudo}systemctl enable ${serverName}
 
   ${csudo}systemctl daemon-reload
-
-  if [ "$verMode" == "cluster" ]; then
-    [ -f ${script_dir}/cfg/nginxd.service ] &&
-      ${csudo}cp ${script_dir}/cfg/nginxd.service \
-        ${service_config_dir}/ || :
-    ${csudo}systemctl daemon-reload
-
-    if ! ${csudo}systemctl enable nginxd &>/dev/null; then
-      ${csudo}systemctl daemon-reexec
-      ${csudo}systemctl enable nginxd
-    fi
-    ${csudo}systemctl start nginxd
-  fi
 }
 
 function install_adapter_service() {
@@ -793,19 +770,6 @@ function updateProduct() {
     sleep 1
   fi
 
-  if [ "$verMode" == "cluster" ]; then
-    if pidof nginx &>/dev/null; then
-      if ((${service_mod} == 0)); then
-        ${csudo}systemctl stop nginxd || :
-      elif ((${service_mod} == 1)); then
-        ${csudo}service nginxd stop || :
-      else
-        kill_process nginx
-      fi
-      sleep 1
-    fi
-  fi
-
   install_main_path
 
   install_log
@@ -817,6 +781,7 @@ function updateProduct() {
   fi
 
   install_examples
+  install_web
   if [ -z $1 ]; then
     install_bin
     install_service
@@ -825,18 +790,6 @@ function updateProduct() {
     install_adapter_config
 
     openresty_work=false
-    if [ "$verMode" == "cluster" ]; then
-      # Check if openresty is installed
-      # Check if nginx is installed successfully
-      if type curl &>/dev/null; then
-        if curl -sSf http://127.0.0.1:${nginx_port} &>/dev/null; then
-          echo -e "\033[44;32;1mNginx for ${productName} is updated successfully!${NC}"
-          openresty_work=true
-        else
-          echo -e "\033[44;31;5mNginx for ${productName} does not work! Please try again!\033[0m"
-        fi
-      fi
-    fi
 
     echo
     echo -e "${GREEN_DARK}To configure ${productName} ${NC}: edit ${cfg_install_dir}/${configFile}"
@@ -857,7 +810,7 @@ function updateProduct() {
     fi
 
     if [ ${openresty_work} = 'true' ]; then
-      echo -e "${GREEN_DARK}To access ${productName}    ${NC}: use ${GREEN_UNDERLINE}${clientName} -h $serverFqdn${NC} in shell OR from ${GREEN_UNDERLINE}http://127.0.0.1:${nginx_port}${NC}"
+      echo -e "${GREEN_DARK}To access ${productName}    ${NC}: use ${GREEN_UNDERLINE}${clientName} -h $serverFqdn${NC} in shell OR from ${GREEN_UNDERLINE}http://127.0.0.1:${web_port}${NC}"
     else
       echo -e "${GREEN_DARK}To access ${productName}    ${NC}: use ${GREEN_UNDERLINE}${clientName} -h $serverFqdn${NC} in shell${NC}"
     fi
@@ -906,6 +859,7 @@ function installProduct() {
       install_connector
   fi
   install_examples
+  install_web
 
   if [ -z $1 ]; then # install service and client
     # For installing new
@@ -915,17 +869,6 @@ function installProduct() {
     install_adapter_config
 
     openresty_work=false
-    if [ "$verMode" == "cluster" ]; then
-      # Check if nginx is installed successfully
-      if type curl &>/dev/null; then
-        if curl -sSf http://127.0.0.1:${nginx_port} &>/dev/null; then
-          echo -e "\033[44;32;1mNginx for ${productName} is installed successfully!${NC}"
-          openresty_work=true
-        else
-          echo -e "\033[44;31;5mNginx for ${productName} does not work! Please try again!\033[0m"
-        fi
-      fi
-    fi
 
     install_config
 

packaging/tools/make_install.sh

@@ -151,6 +151,7 @@ function install_main_path() {
     ${csudo}mkdir -p ${install_main_dir}/driver
     ${csudo}mkdir -p ${install_main_dir}/examples
     ${csudo}mkdir -p ${install_main_dir}/include
+    ${csudo}mkdir -p ${install_main_dir}/share
     #    ${csudo}mkdir -p ${install_main_dir}/init.d
   else
     ${csudo}rm -rf ${install_main_dir} || ${csudo}rm -rf ${install_main_2_dir} || :
@@ -161,6 +162,7 @@ function install_main_path() {
     ${csudo}mkdir -p ${install_main_dir}/driver || ${csudo}mkdir -p ${install_main_2_dir}/driver
     ${csudo}mkdir -p ${install_main_dir}/examples || ${csudo}mkdir -p ${install_main_2_dir}/examples
     ${csudo}mkdir -p ${install_main_dir}/include || ${csudo}mkdir -p ${install_main_2_dir}/include
+    ${csudo}mkdir -p ${install_main_dir}/share || ${csudo}mkdir -p ${install_main_2_dir}/share
   fi
 }
 
@@ -469,6 +471,16 @@ function install_examples() {
   fi
 }
 
+function install_web() {
+  if [ -d "${binary_dir}/build/share" ]; then
+    if [ "$osType" != "Darwin" ]; then
+      ${csudo}cp -rf ${binary_dir}/build/share/* ${install_main_dir}/share || :
+    else
+      ${csudo}cp -rf ${binary_dir}/build/share/* ${install_main_dir}/share || ${csudo}cp -rf ${binary_dir}/build/share/* ${install_main_2_dir}/share || :
+    fi
+  fi
+}
+
 function clean_service_on_sysvinit() {
   if pidof ${serverName} &>/dev/null; then
     ${csudo}service ${serverName} stop || :
@@ -596,6 +608,7 @@ function update_TDengine() {
   install_lib
   #  install_connector
   install_examples
+  install_web
   install_bin
 
   install_service

packaging/tools/makepkg.sh

@@ -107,7 +107,7 @@ else
 fi
 
 install_files="${script_dir}/install.sh"
-nginx_dir="${top_dir}/../enterprise/src/plugins/web"
+web_dir="${top_dir}/../enterprise/src/plugins/web"
 
 init_file_deb=${script_dir}/../deb/taosd
 init_file_rpm=${script_dir}/../rpm/taosd
@@ -132,10 +132,6 @@ if [ -f "${cfg_dir}/${serverName}.service" ]; then
   cp ${cfg_dir}/${serverName}.service ${install_dir}/cfg || :
 fi
 
-if [ -f "${top_dir}/packaging/cfg/nginxd.service" ]; then
-  cp ${top_dir}/packaging/cfg/nginxd.service ${install_dir}/cfg || :
-fi
-
 mkdir -p ${install_dir}/bin && cp ${bin_files} ${install_dir}/bin && chmod a+x ${install_dir}/bin/* || :
 mkdir -p ${install_dir}/init.d && cp ${init_file_deb} ${install_dir}/init.d/${serverName}.deb
 mkdir -p ${install_dir}/init.d && cp ${init_file_rpm} ${install_dir}/init.d/${serverName}.rpm
@@ -222,16 +218,6 @@ if [ "$verMode" == "cluster" ]; then
   sed 's/verMode=edge/verMode=cluster/g' ${install_dir}/bin/remove.sh >>remove_temp.sh
   mv remove_temp.sh ${install_dir}/bin/remove.sh
 
-  mkdir -p ${install_dir}/nginxd && cp -r ${nginx_dir}/* ${install_dir}/nginxd
-  cp ${nginx_dir}/png/taos.png ${install_dir}/nginxd/admin/images/taos.png
-  rm -rf ${install_dir}/nginxd/png
-
-  if [ "$cpuType" == "aarch64" ]; then
-    cp -f ${install_dir}/nginxd/sbin/arm/64bit/nginx ${install_dir}/nginxd/sbin/
-  elif [ "$cpuType" == "aarch32" ]; then
-    cp -f ${install_dir}/nginxd/sbin/arm/32bit/nginx ${install_dir}/nginxd/sbin/
-  fi
-  rm -rf ${install_dir}/nginxd/sbin/arm
 fi
 
 cd ${install_dir}
@@ -288,6 +274,13 @@ if [[ $dbName == "taos" ]]; then
     cp -r ${examples_dir}/C# ${install_dir}/examples
     mkdir -p ${install_dir}/examples/taosbenchmark-json && cp ${examples_dir}/../tools/taos-tools/example/* ${install_dir}/examples/taosbenchmark-json
   fi
+
+  # Add web files
+  if [ -d "${web_dir}/admin" ]; then
+    mkdir -p ${install_dir}/share/
+    cp ${web_dir}/admin ${install_dir}/share/ -r
+    cp ${web_dir}/png/taos.png ${install_dir}/share/admin/images/taos.png
+  fi
 fi
 
 # Copy driver

packaging/tools/remove.sh

@@ -27,13 +27,11 @@ local_bin_link_dir="/usr/local/bin"
 lib_link_dir="/usr/lib"
 lib64_link_dir="/usr/lib64"
 inc_link_dir="/usr/include"
-install_nginxd_dir="/usr/local/nginxd"
 
 service_config_dir="/etc/systemd/system"
 taos_service_name=${serverName}
 taosadapter_service_name="taosadapter"
 tarbitrator_service_name="tarbitratord"
-nginx_service_name="nginxd"
 csudo=""
 if command -v sudo >/dev/null; then
   csudo="sudo "
@@ -153,18 +151,6 @@ function clean_service_on_systemd() {
   fi
   ${csudo}systemctl disable ${tarbitrator_service_name} &>/dev/null || echo &>/dev/null
   ${csudo}rm -f ${tarbitratord_service_config}
-
-  if [ "$verMode" == "cluster" ]; then
-    nginx_service_config="${service_config_dir}/${nginx_service_name}.service"
-    if [ -d ${install_nginxd_dir} ]; then
-      if systemctl is-active --quiet ${nginx_service_name}; then
-        echo "Nginx for ${productName} is running, stopping it..."
-        ${csudo}systemctl stop ${nginx_service_name} &>/dev/null || echo &>/dev/null
-      fi
-      ${csudo}systemctl disable ${nginx_service_name} &>/dev/null || echo &>/dev/null
-      ${csudo}rm -f ${nginx_service_config}
-    fi
-  fi
 }
 
 function clean_service_on_sysvinit() {
@@ -239,7 +225,6 @@ clean_config
 ${csudo}rm -rf ${data_link_dir} || :
 
 ${csudo}rm -rf ${install_main_dir}
-${csudo}rm -rf ${install_nginxd_dir}
 if [[ -e /etc/os-release ]]; then
   osinfo=$(awk -F= '/^NAME/{print $2}' /etc/os-release)
 else

source/client/src/clientEnv.c

@@ -19,6 +19,7 @@
 #include "functionMgt.h"
 #include "os.h"
 #include "query.h"
+#include "qworker.h"
 #include "scheduler.h"
 #include "tcache.h"
 #include "tglobal.h"
@@ -27,7 +28,6 @@
 #include "trpc.h"
 #include "tsched.h"
 #include "ttime.h"
-#include "qworker.h"
 
 #define TSC_VAR_NOT_RELEASE 1
 #define TSC_VAR_RELEASED    0
@@ -70,11 +70,10 @@ static void deregisterRequest(SRequestObj *pRequest) {
   int32_t currentInst = atomic_sub_fetch_64((int64_t *)&pActivity->currentRequests, 1);
   int32_t num = atomic_sub_fetch_32(&pTscObj->numOfReqs, 1);
 
-  int64_t nowUs = taosGetTimestampUs();
-  int64_t duration = nowUs - pRequest->metric.start;
-  tscDebug("0x%" PRIx64 " free Request from connObj: 0x%" PRIx64 ", reqId:0x%" PRIx64 " elapsed:%" PRIu64
-           " ms, current:%d, app current:%d",
-           pRequest->self, pTscObj->id, pRequest->requestId, duration / 1000, num, currentInst);
+  int64_t duration = taosGetTimestampUs() - pRequest->metric.start;
+  tscDebug("0x%" PRIx64 " free Request from connObj: 0x%" PRIx64 ", reqId:0x%" PRIx64 " elapsed:%.2f ms, "
+           "current:%d, app current:%d",
+           pRequest->self, pTscObj->id, pRequest->requestId, duration / 1000.0, num, currentInst);
 
   if (QUERY_NODE_VNODE_MODIF_STMT == pRequest->stmtType) {
     tscPerf("insert duration %" PRId64 "us: syntax:%" PRId64 "us, ctg:%" PRId64 "us, semantic:%" PRId64
@@ -85,11 +84,12 @@ static void deregisterRequest(SRequestObj *pRequest) {
     atomic_add_fetch_64((int64_t *)&pActivity->insertElapsedTime, duration);
   } else if (QUERY_NODE_SELECT_STMT == pRequest->stmtType) {
     tscPerf("select duration %" PRId64 "us: syntax:%" PRId64 "us, ctg:%" PRId64 "us, semantic:%" PRId64
-            "us, planner:%" PRId64 "us, exec:%" PRId64 "us",
+            "us, planner:%" PRId64 "us, exec:%" PRId64 "us, reqId:0x%"PRIx64,
             duration, pRequest->metric.syntaxEnd - pRequest->metric.syntaxStart,
             pRequest->metric.ctgEnd - pRequest->metric.ctgStart, pRequest->metric.semanticEnd - pRequest->metric.ctgEnd,
             pRequest->metric.planEnd - pRequest->metric.semanticEnd,
-            pRequest->metric.resultReady - pRequest->metric.planEnd);
+            pRequest->metric.resultReady - pRequest->metric.planEnd, pRequest->requestId);
+
     atomic_add_fetch_64((int64_t *)&pActivity->queryElapsedTime, duration);
   }
 

source/client/src/clientImpl.c

@@ -483,8 +483,8 @@ void setResPrecision(SReqResultInfo* pResInfo, int32_t precision) {
 
 int32_t buildVnodePolicyNodeList(SRequestObj* pRequest, SArray** pNodeList, SArray* pMnodeList, SArray* pDbVgList) {
   SArray* nodeList = taosArrayInit(4, sizeof(SQueryNodeLoad));
-  char *policy = (tsQueryPolicy == QUERY_POLICY_VNODE) ? "vnode" : "client";
-  
+  char*   policy = (tsQueryPolicy == QUERY_POLICY_VNODE) ? "vnode" : "client";
+
   int32_t dbNum = taosArrayGetSize(pDbVgList);
   for (int32_t i = 0; i < dbNum; ++i) {
     SArray* pVg = taosArrayGetP(pDbVgList, i);
@@ -815,7 +815,7 @@ int32_t handleCreateTbExecRes(void* res, SCatalog* pCatalog) {
 
 int32_t handleQueryExecRsp(SRequestObj* pRequest) {
   if (NULL == pRequest->body.resInfo.execRes.res) {
-    return TSDB_CODE_SUCCESS;
+    return pRequest->code;
   }
 
   SCatalog*     pCatalog = NULL;
@@ -868,10 +868,12 @@ int32_t handleQueryExecRsp(SRequestObj* pRequest) {
   return code;
 }
 
+//todo refacto the error code  mgmt
 void schedulerExecCb(SExecResult* pResult, void* param, int32_t code) {
   SRequestObj* pRequest = (SRequestObj*)param;
-  pRequest->code = code;
+  STscObj* pTscObj = pRequest->pTscObj;
 
+  pRequest->code = code;
   pRequest->metric.resultReady = taosGetTimestampUs();
 
   if (pResult) {
@@ -879,33 +881,32 @@ void schedulerExecCb(SExecResult* pResult, void* param, int32_t code) {
     memcpy(&pRequest->body.resInfo.execRes, pResult, sizeof(*pResult));
   }
 
-  if (TDMT_VND_SUBMIT == pRequest->type || TDMT_VND_DELETE == pRequest->type ||
-      TDMT_VND_CREATE_TABLE == pRequest->type) {
+  int32_t type = pRequest->type;
+  if (TDMT_VND_SUBMIT == type || TDMT_VND_DELETE == type || TDMT_VND_CREATE_TABLE == type) {
     if (pResult) {
       pRequest->body.resInfo.numOfRows = pResult->numOfRows;
-      if (TDMT_VND_SUBMIT == pRequest->type) {
-        STscObj*            pTscObj = pRequest->pTscObj;
+
+      // record the insert rows
+      if (TDMT_VND_SUBMIT == type) {
         SAppClusterSummary* pActivity = &pTscObj->pAppInfo->summary;
         atomic_add_fetch_64((int64_t*)&pActivity->numOfInsertRows, pResult->numOfRows);
       }
     }
 
     schedulerFreeJob(&pRequest->body.queryJob, 0);
-
-    pRequest->metric.execEnd = taosGetTimestampUs();
   }
 
   taosMemoryFree(pResult);
+  tscDebug("0x%" PRIx64 " enter scheduler exec cb, code:%s, reqId:0x%" PRIx64, pRequest->self, tstrerror(code),
+           pRequest->requestId);
 
-  tscDebug("0x%" PRIx64 " enter scheduler exec cb, code:%d - %s, reqId:0x%" PRIx64, pRequest->self, code,
-           tstrerror(code), pRequest->requestId);
-
-  STscObj* pTscObj = pRequest->pTscObj;
   if (code != TSDB_CODE_SUCCESS && NEED_CLIENT_HANDLE_ERROR(code) && pRequest->sqlstr != NULL) {
-    tscDebug("0x%" PRIx64 " client retry to handle the error, code:%d - %s, tryCount:%d, reqId:0x%" PRIx64,
-             pRequest->self, code, tstrerror(code), pRequest->retry, pRequest->requestId);
+    tscDebug("0x%" PRIx64 " client retry to handle the error, code:%s, tryCount:%d, reqId:0x%" PRIx64,
+             pRequest->self, tstrerror(code), pRequest->retry, pRequest->requestId);
     pRequest->prevCode = code;
     schedulerFreeJob(&pRequest->body.queryJob, 0);
+    qDestroyQuery(pRequest->pQuery);
+    pRequest->pQuery = NULL;
     doAsyncQuery(pRequest, true);
     return;
   }
@@ -915,7 +916,11 @@ void schedulerExecCb(SExecResult* pResult, void* param, int32_t code) {
     removeMeta(pTscObj, pRequest->targetTableList);
   }
 
-  handleQueryExecRsp(pRequest);
+  pRequest->metric.execEnd = taosGetTimestampUs();
+  int32_t code1 = handleQueryExecRsp(pRequest);
+  if (pRequest->code == TSDB_CODE_SUCCESS && pRequest->code != code1) {
+    pRequest->code = code1;
+  }
 
   // return to client
   pRequest->body.queryFp(pRequest->body.param, pRequest, code);

source/client/src/clientMain.c

@@ -20,13 +20,13 @@
 #include "functionMgt.h"
 #include "os.h"
 #include "query.h"
+#include "qworker.h"
 #include "scheduler.h"
 #include "tglobal.h"
 #include "tmsg.h"
 #include "tref.h"
 #include "trpc.h"
 #include "version.h"
-#include "qworker.h"
 
 #define TSC_VAR_NOT_RELEASE 1
 #define TSC_VAR_RELEASED    0
@@ -700,6 +700,7 @@ void retrieveMetaCallback(SMetaData *pResultMeta, void *param, int32_t code) {
   SQuery          *pQuery = pRequest->pQuery;
 
   pRequest->metric.ctgEnd = taosGetTimestampUs();
+  qDebug("0x%" PRIx64 " start to semantic analysis, reqId:0x%" PRIx64, pRequest->self, pRequest->requestId);
 
   if (code == TSDB_CODE_SUCCESS) {
     code = qAnalyseSqlSemantic(pWrapper->pCtx, &pWrapper->catalogReq, pResultMeta, pQuery);
@@ -723,13 +724,16 @@ void retrieveMetaCallback(SMetaData *pResultMeta, void *param, int32_t code) {
 
     destorySqlParseWrapper(pWrapper);
 
-    tscDebug("0x%" PRIx64 " analysis semantics completed, start async query, reqId:0x%" PRIx64, pRequest->self,
-             pRequest->requestId);
+    double el = (pRequest->metric.semanticEnd - pRequest->metric.ctgEnd)/1000.0;
+    tscDebug("0x%" PRIx64 " analysis semantics completed, start async query, elapsed time:%.2f ms, reqId:0x%" PRIx64,
+             pRequest->self, el, pRequest->requestId);
+
     launchAsyncQuery(pRequest, pQuery, pResultMeta);
   } else {
     destorySqlParseWrapper(pWrapper);
     qDestroyQuery(pRequest->pQuery);
     pRequest->pQuery = NULL;
+
     if (NEED_CLIENT_HANDLE_ERROR(code)) {
       tscDebug("0x%" PRIx64 " client retry to handle the error, code:%d - %s, tryCount:%d, reqId:0x%" PRIx64,
                pRequest->self, code, tstrerror(code), pRequest->retry, pRequest->requestId);

source/common/src/tglobal.c

@@ -85,8 +85,7 @@ uint16_t tsTelemPort = 80;
 char tsSmlTagName[TSDB_COL_NAME_LEN] = "_tag_null";
 char tsSmlChildTableName[TSDB_TABLE_NAME_LEN] = "";  // user defined child table name can be specified in tag value.
                                                      // If set to empty system will generate table name using MD5 hash.
-bool tsSmlDataFormat =
-    true;  // true means that the name and order of cols in each line are the same(only for influx protocol)
+bool tsSmlDataFormat = false;  // true means that the name and order of cols in each line are the same(only for influx protocol)
 
 // query
 int32_t tsQueryPolicy = 1;
@@ -95,6 +94,7 @@ int32_t tsQueryRsmaTolerance = 1000;  // the tolerance time (ms) to judge from w
 bool    tsQueryPlannerTrace = false;
 int32_t tsQueryNodeChunkSize = 32 * 1024;
 bool    tsQueryUseNodeAllocator = true;
+bool    tsKeepColumnName = false;
 
 /*
  * denote if the server needs to compress response message at the application layer to client, including query rsp,
@@ -205,7 +205,9 @@ static int32_t taosLoadCfg(SConfig *pCfg, const char **envCmd, const char *input
     tstrncpy(cfgFile, cfgDir, sizeof(cfgDir));
   }
 
-  if (apolloUrl == NULL || apolloUrl[0] == '\0') cfgGetApollUrl(envCmd, envFile, apolloUrl);
+  if (apolloUrl != NULL && apolloUrl[0] == '\0') {
+    cfgGetApollUrl(envCmd, envFile, apolloUrl);
+  }
 
   if (cfgLoad(pCfg, CFG_STYPE_APOLLO_URL, apolloUrl) != 0) {
     uError("failed to load from apollo url:%s since %s", apolloUrl, terrstr());
@@ -290,6 +292,7 @@ static int32_t taosAddClientCfg(SConfig *pCfg) {
   if (cfgAddBool(pCfg, "queryPlannerTrace", tsQueryPlannerTrace, true) != 0) return -1;
   if (cfgAddInt32(pCfg, "queryNodeChunkSize", tsQueryNodeChunkSize, 1024, 128 * 1024, true) != 0) return -1;
   if (cfgAddBool(pCfg, "queryUseNodeAllocator", tsQueryUseNodeAllocator, true) != 0) return -1;
+  if (cfgAddBool(pCfg, "keepColumnName", tsKeepColumnName, true) != 0) return -1;
   if (cfgAddString(pCfg, "smlChildTableName", "", 1) != 0) return -1;
   if (cfgAddString(pCfg, "smlTagName", tsSmlTagName, 1) != 0) return -1;
   if (cfgAddBool(pCfg, "smlDataFormat", tsSmlDataFormat, 1) != 0) return -1;
@@ -652,6 +655,7 @@ static int32_t taosSetClientCfg(SConfig *pCfg) {
   tsQueryPlannerTrace = cfgGetItem(pCfg, "queryPlannerTrace")->bval;
   tsQueryNodeChunkSize = cfgGetItem(pCfg, "queryNodeChunkSize")->i32;
   tsQueryUseNodeAllocator = cfgGetItem(pCfg, "queryUseNodeAllocator")->bval;
+  tsKeepColumnName = cfgGetItem(pCfg, "keepColumnName")->bval;
   return 0;
 }
 
@@ -845,6 +849,9 @@ int32_t taosSetCfg(SConfig *pCfg, char *name) {
       break;
     }
     case 'k': {
+      if (strcasecmp("keepColumnName", name) == 0) {
+        tsKeepColumnName = cfgGetItem(pCfg, "keepColumnName")->bval;
+      }
       break;
     }
     case 'l': {
@@ -921,9 +928,9 @@ int32_t taosSetCfg(SConfig *pCfg, char *name) {
         }
         case 'u': {
           if (strcasecmp("multiProcess", name) == 0) {
-          #if !defined(WINDOWS) && !defined(DARWIN)
+#if !defined(WINDOWS) && !defined(DARWIN)
             tsMultiProcess = cfgGetItem(pCfg, "multiProcess")->bval;
-          #endif
+#endif
           } else if (strcasecmp("udfDebugFlag", name) == 0) {
             udfDebugFlag = cfgGetItem(pCfg, "udfDebugFlag")->i32;
           }
@@ -1126,11 +1133,20 @@ int32_t taosCreateLog(const char *logname, int32_t logFileNum, const char *cfgDi
 
   if (tsc) {
     tsLogEmbedded = 0;
-    if (taosAddClientLogCfg(pCfg) != 0) return -1;
+    if (taosAddClientLogCfg(pCfg) != 0) {
+      cfgCleanup(pCfg);
+      return -1;
+    }
   } else {
     tsLogEmbedded = 1;
-    if (taosAddClientLogCfg(pCfg) != 0) return -1;
-    if (taosAddServerLogCfg(pCfg) != 0) return -1;
+    if (taosAddClientLogCfg(pCfg) != 0) {
+      cfgCleanup(pCfg);
+      return -1;
+    }
+    if (taosAddServerLogCfg(pCfg) != 0) {
+      cfgCleanup(pCfg);
+      return -1;
+    }
   }
 
   if (taosLoadCfg(pCfg, envCmd, cfgDir, envFile, apolloUrl) != 0) {

source/dnode/mgmt/mgmt_qnode/src/qmWorker.c

@@ -87,6 +87,7 @@ int32_t qmPutRpcMsgToQueue(SQnodeMgmt *pMgmt, EQueueType qtype, SRpcMsg *pRpc) {
       return 0;
     default:
       terrno = TSDB_CODE_INVALID_PARA;
+      taosFreeQitem(pMsg);
       return -1;
   }
 }

source/dnode/mgmt/mgmt_vnode/src/vmFile.c

@@ -135,12 +135,14 @@ _OVER:
   if (content != NULL) taosMemoryFree(content);
   if (root != NULL) cJSON_Delete(root);
   if (pFile != NULL) taosCloseFile(&pFile);
+  if (*ppCfgs == NULL && pCfgs != NULL) taosMemoryFree(pCfgs);
 
   terrno = code;
   return code;
 }
 
 int32_t vmWriteVnodeListToFile(SVnodeMgmt *pMgmt) {
+  int32_t ret = 0;
   char file[PATH_MAX] = {0};
   char realfile[PATH_MAX] = {0};
   snprintf(file, sizeof(file), "%s%svnodes.json.bak", pMgmt->path, TD_DIRSEP);
@@ -161,13 +163,16 @@ int32_t vmWriteVnodeListToFile(SVnodeMgmt *pMgmt) {
   char   *content = taosMemoryCalloc(1, maxLen + 1);
   if (content == NULL) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
-    return -1;
+    ret = -1;
+    goto _OVER;
   }
 
   len += snprintf(content + len, maxLen - len, "{\n");
   len += snprintf(content + len, maxLen - len, "  \"vnodes\": [\n");
   for (int32_t i = 0; i < numOfVnodes; ++i) {
     SVnodeObj *pVnode = pVnodes[i];
+    if (pVnode == NULL) continue;
+
     len += snprintf(content + len, maxLen - len, "    {\n");
     len += snprintf(content + len, maxLen - len, "      \"vgId\": %d,\n", pVnode->vgId);
     len += snprintf(content + len, maxLen - len, "      \"dropped\": %d,\n", pVnode->dropped);
@@ -180,12 +185,13 @@ int32_t vmWriteVnodeListToFile(SVnodeMgmt *pMgmt) {
   }
   len += snprintf(content + len, maxLen - len, "  ]\n");
   len += snprintf(content + len, maxLen - len, "}\n");
+  terrno = 0;
 
+_OVER:
   taosWriteFile(pFile, content, len);
   taosFsyncFile(pFile);
   taosCloseFile(&pFile);
   taosMemoryFree(content);
-  terrno = 0;
 
   for (int32_t i = 0; i < numOfVnodes; ++i) {
     SVnodeObj *pVnode = pVnodes[i];
@@ -196,6 +202,8 @@ int32_t vmWriteVnodeListToFile(SVnodeMgmt *pMgmt) {
     taosMemoryFree(pVnodes);
   }
 
+  if (ret != 0) return -1;
+
   dDebug("successed to write %s, numOfVnodes:%d", realfile, numOfVnodes);
   return taosRenameFile(file, realfile);
 }
\ No newline at end of file

source/dnode/mgmt/mgmt_vnode/src/vmHandle.c

@@ -175,7 +175,7 @@ static void vmGenerateVnodeCfg(SCreateVnodeReq *pCreate, SVnodeCfg *pCfg) {
   pCfg->hashSuffix = pCreate->hashSuffix;
   pCfg->tsdbPageSize = pCreate->tsdbPageSize * 1024;
 
-  pCfg->standby = pCfg->standby;
+  pCfg->standby = 0;
   pCfg->syncCfg.myIndex = pCreate->selfIndex;
   pCfg->syncCfg.replicaNum = pCreate->replica;
   memset(&pCfg->syncCfg.nodeInfo, 0, sizeof(pCfg->syncCfg.nodeInfo));

source/dnode/mgmt/mgmt_vnode/src/vmInt.c

@@ -58,11 +58,14 @@ int32_t vmOpenVnode(SVnodeMgmt *pMgmt, SWrapperCfg *pCfg, SVnode *pImpl) {
 
   if (pVnode->path == NULL) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
+    taosMemoryFree(pVnode);
     return -1;
   }
 
   if (vmAllocQueue(pMgmt, pVnode) != 0) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
+    taosMemoryFree(pVnode->path);
+    taosMemoryFree(pVnode);
     return -1;
   }
 
@@ -221,6 +224,7 @@ static void vmCloseVnodes(SVnodeMgmt *pMgmt) {
   SVnodeObj **ppVnodes = vmGetVnodeListFromHash(pMgmt, &numOfVnodes);
 
   for (int32_t i = 0; i < numOfVnodes; ++i) {
+    if (ppVnodes == NULL || ppVnodes[i] == NULL) continue;
     vmCloseVnode(pMgmt, ppVnodes[i]);
   }
 
@@ -380,7 +384,9 @@ static int32_t vmStartVnodes(SVnodeMgmt *pMgmt) {
   for (int32_t v = 0; v < numOfVnodes; ++v) {
     int32_t       t = v % threadNum;
     SVnodeThread *pThread = &threads[t];
-    pThread->ppVnodes[pThread->vnodeNum++] = ppVnodes[v];
+    if (pThread->ppVnodes != NULL) {
+      pThread->ppVnodes[pThread->vnodeNum++] = ppVnodes[v];
+    }
   }
 
   pMgmt->state.openVnodes = 0;
@@ -411,8 +417,8 @@ static int32_t vmStartVnodes(SVnodeMgmt *pMgmt) {
   taosMemoryFree(threads);
 
   for (int32_t i = 0; i < numOfVnodes; ++i) {
-    SVnodeObj *pVnode = ppVnodes[i];
-    vmReleaseVnode(pMgmt, pVnode);
+    if (ppVnodes == NULL || ppVnodes[i] == NULL) continue;
+    vmReleaseVnode(pMgmt, ppVnodes[i]);
   }
 
   if (ppVnodes != NULL) {

source/dnode/mnode/sdb/src/sdbFile.c

@@ -514,7 +514,7 @@ static void sdbCloseIter(SSdbIter *pIter) {
   }
 
   if (pIter->name != NULL) {
-    taosRemoveFile(pIter->name);
+    (void)taosRemoveFile(pIter->name);
     taosMemoryFree(pIter->name);
     pIter->name = NULL;
   }
@@ -606,6 +606,7 @@ int32_t sdbStartWrite(SSdb *pSdb, SSdbIter **ppIter) {
   if (pIter->file == NULL) {
     terrno = TAOS_SYSTEM_ERROR(errno);
     mError("failed to open %s since %s", pIter->name, terrstr());
+    sdbCloseIter(pIter);
     return -1;
   }
 
@@ -636,9 +637,9 @@ int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply, int64_t index, i
     return -1;
   }
 
-  sdbCloseIter(pIter);
   if (sdbReadFile(pSdb) != 0) {
     mError("sdbiter:%p, failed to read from %s since %s", pIter, datafile, terrstr());
+    sdbCloseIter(pIter);
     return -1;
   }
 
@@ -656,6 +657,7 @@ int32_t sdbStopWrite(SSdb *pSdb, SSdbIter *pIter, bool isApply, int64_t index, i
   }
 
   mInfo("sdbiter:%p, success applyed to sdb", pIter);
+  sdbCloseIter(pIter);
   return 0;
 }
 

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

@@ -32,6 +32,12 @@ extern "C" {
 #define tsdbTrace(...) do { if (tsdbDebugFlag & DEBUG_TRACE) { taosPrintLog("TSD ", DEBUG_TRACE, tsdbDebugFlag, __VA_ARGS__); }} while(0)
 // clang-format on
 
+#define TSDB_CHECK_CODE(CODE, LINO, LABEL) \
+  if (CODE) {                              \
+    LINO = __LINE__;                       \
+    goto LABEL;                            \
+  }
+
 typedef struct TSDBROW       TSDBROW;
 typedef struct TABLEID       TABLEID;
 typedef struct TSDBKEY       TSDBKEY;
@@ -150,7 +156,7 @@ int32_t tCmprBlockL(void const *lhs, void const *rhs);
 
 int32_t   tBlockDataCreate(SBlockData *pBlockData);
 void      tBlockDataDestroy(SBlockData *pBlockData, int8_t deepClear);
-int32_t   tBlockDataInit(SBlockData *pBlockData, int64_t suid, int64_t uid, STSchema *pTSchema);
+int32_t   tBlockDataInit(SBlockData *pBlockData, TABLEID *pId, STSchema *pTSchema, int16_t *aCid, int32_t nCid);
 int32_t   tBlockDataInitEx(SBlockData *pBlockData, SBlockData *pBlockDataFrom);
 void      tBlockDataReset(SBlockData *pBlockData);
 int32_t   tBlockDataAppendRow(SBlockData *pBlockData, TSDBROW *pRow, STSchema *pTSchema, int64_t uid);
@@ -272,6 +278,7 @@ int32_t tsdbReadSttBlk(SDataFReader *pReader, int32_t iStt, SArray *aSttBlk);
 int32_t tsdbReadBlockSma(SDataFReader *pReader, SDataBlk *pBlock, SArray *aColumnDataAgg);
 int32_t tsdbReadDataBlock(SDataFReader *pReader, SDataBlk *pBlock, SBlockData *pBlockData);
 int32_t tsdbReadSttBlock(SDataFReader *pReader, int32_t iStt, SSttBlk *pSttBlk, SBlockData *pBlockData);
+int32_t tsdbReadSttBlockEx(SDataFReader *pReader, int32_t iStt, SSttBlk *pSttBlk, SBlockData *pBlockData);
 // SDelFWriter
 int32_t tsdbDelFWriterOpen(SDelFWriter **ppWriter, SDelFile *pFile, STsdb *pTsdb);
 int32_t tsdbDelFWriterClose(SDelFWriter **ppWriter, int8_t sync);
@@ -633,6 +640,9 @@ typedef struct SSttBlockLoadInfo {
   int32_t    currentLoadBlockIndex;
   int32_t    loadBlocks;
   double     elapsedTime;
+  STSchema  *pSchema;
+  int16_t   *colIds;
+  int32_t    numOfCols;
 } SSttBlockLoadInfo;
 
 typedef struct SMergeTree {
@@ -652,13 +662,14 @@ typedef struct {
 } SSkmInfo;
 
 int32_t tMergeTreeOpen(SMergeTree *pMTree, int8_t backward, SDataFReader *pFReader, uint64_t suid, uint64_t uid,
-                       STimeWindow *pTimeWindow, SVersionRange *pVerRange, void *pLoadInfo, const char *idStr);
+                       STimeWindow *pTimeWindow, SVersionRange *pVerRange, SSttBlockLoadInfo *pBlockLoadInfo,
+                       bool destroyLoadInfo, const char *idStr);
 void    tMergeTreeAddIter(SMergeTree *pMTree, SLDataIter *pIter);
 bool    tMergeTreeNext(SMergeTree *pMTree);
 TSDBROW tMergeTreeGetRow(SMergeTree *pMTree);
 void    tMergeTreeClose(SMergeTree *pMTree);
 
-SSttBlockLoadInfo *tCreateLastBlockLoadInfo();
+SSttBlockLoadInfo *tCreateLastBlockLoadInfo(STSchema *pSchema, int16_t *colList, int32_t numOfCols);
 void               resetLastBlockLoadInfo(SSttBlockLoadInfo *pLoadInfo);
 void               getLastBlockLoadInfo(SSttBlockLoadInfo *pLoadInfo, int64_t *blocks, double *el);
 void              *destroyLastBlockLoadInfo(SSttBlockLoadInfo *pLoadInfo);

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

@@ -872,7 +872,7 @@ SArray *metaGetSmaIdsByTable(SMeta *pMeta, tb_uid_t uid) {
 
     pSmaIdxKey = (SSmaIdxKey *)pCur->pKey;
 
-    if (taosArrayPush(pUids, &pSmaIdxKey->smaUid) < 0) {
+    if (!taosArrayPush(pUids, &pSmaIdxKey->smaUid)) {
       terrno = TSDB_CODE_OUT_OF_MEMORY;
       metaCloseSmaCursor(pCur);
       taosArrayDestroy(pUids);
@@ -915,7 +915,7 @@ SArray *metaGetSmaTbUids(SMeta *pMeta) {
       }
     }
 
-    if (taosArrayPush(pUids, &uid) < 0) {
+    if (!taosArrayPush(pUids, &uid)) {
       terrno = TSDB_CODE_OUT_OF_MEMORY;
       metaCloseSmaCursor(pCur);
       taosArrayDestroy(pUids);

source/dnode/vnode/src/sma/smaCommit.c

@@ -213,7 +213,7 @@ static int32_t tdUpdateQTaskInfoFiles(SSma *pSma, SRSmaStat *pStat) {
     tdRSmaQTaskInfoGetFullName(TD_VID(pVnode), committed, tfsGetPrimaryPath(pVnode->pTfs), qTaskInfoFullName);
     if (taosCheckExistFile(qTaskInfoFullName)) {
       SQTaskFile qFile = {.nRef = 1, .padding = 0, .version = committed, .size = 0};
-      if (taosArrayPush(pFS->aQTaskInf, &qFile) < 0) {
+      if (!taosArrayPush(pFS->aQTaskInf, &qFile)) {
         taosWUnLockLatch(RSMA_FS_LOCK(pStat));
         terrno = TSDB_CODE_OUT_OF_MEMORY;
         return TSDB_CODE_FAILED;

source/dnode/vnode/src/sma/smaEnv.c

@@ -386,7 +386,7 @@ int32_t tdCheckAndInitSmaEnv(SSma *pSma, int8_t smaType) {
       }
       break;
     default:
-      smaError("vgId:%d, undefined smaType:%", SMA_VID(pSma), smaType);
+      smaError("vgId:%d, undefined smaType:%" PRIi8, SMA_VID(pSma), smaType);
       return TSDB_CODE_FAILED;
   }
 

source/dnode/vnode/src/sma/smaFS.c

@@ -54,7 +54,7 @@ int32_t tdRSmaFSOpen(SSma *pSma, int64_t version) {
     if ((terrno = tdRSmaFSUpsertQTaskFile(RSMA_FS(pStat), &qTaskFile)) < 0) {
       goto _end;
     }
-    smaInfo("vgId:%d, open fs, version:%" PRIi64 ", ref:%" PRIi64, TD_VID(pVnode), qTaskFile.version, qTaskFile.nRef);
+    smaInfo("vgId:%d, open fs, version:%" PRIi64 ", ref:%d", TD_VID(pVnode), qTaskFile.version, qTaskFile.nRef);
   }
 
 _end:

source/dnode/vnode/src/sma/smaRollup.c

@@ -196,7 +196,8 @@ static int32_t tdUpdateTbUidListImpl(SSma *pSma, tb_uid_t *suid, SArray *tbUids,
 
   if (!suid || !tbUids) {
     terrno = TSDB_CODE_INVALID_PTR;
-    smaError("vgId:%d, failed to get rsma info for uid:%" PRIi64 " since %s", SMA_VID(pSma), *suid, terrstr());
+    smaError("vgId:%d, failed to get rsma info for uid:%" PRIi64 " since %s", SMA_VID(pSma), suid ? *suid : -1,
+             terrstr());
     return TSDB_CODE_FAILED;
   }
 
@@ -566,6 +567,7 @@ static int32_t tdUidStorePut(STbUidStore *pStore, tb_uid_t suid, tb_uid_t *uid)
         }
         if (!taosArrayPush(pUidArray, uid)) {
           terrno = TSDB_CODE_OUT_OF_MEMORY;
+          taosArrayDestroy(pUidArray);
           return TSDB_CODE_FAILED;
         }
         if (taosHashPut(pStore->uidHash, &suid, sizeof(suid), &pUidArray, sizeof(pUidArray)) < 0) {
@@ -690,7 +692,8 @@ static int32_t tdRSmaExecAndSubmitResult(SSma *pSma, qTaskInfo_t taskInfo, SRSma
 
   while (1) {
     uint64_t ts;
-    int32_t  code = qExecTaskOpt(taskInfo, pResList, &ts, NULL);
+    bool hasMore = false;
+    int32_t  code = qExecTaskOpt(taskInfo, pResList, &ts, &hasMore, NULL);
     if (code < 0) {
       if (code == TSDB_CODE_QRY_IN_EXEC) {
         break;
@@ -1677,13 +1680,13 @@ static void tdRSmaFetchTrigger(void *param, void *tmrId) {
   SRSmaInfoItem *pItem = NULL;
 
   if (!(pRSmaRef = taosHashGet(smaMgmt.refHash, &param, POINTER_BYTES))) {
-    smaDebug("rsma fetch task not start since rsma info item:%p not exist in refHash:%p, rsetId:%" PRIi64, param,
-             *(int64_t *)&param, smaMgmt.refHash, smaMgmt.rsetId);
+    smaDebug("rsma fetch task not start since rsma info item:%p not exist in refHash:%p, rsetId:%d", param,
+             smaMgmt.refHash, smaMgmt.rsetId);
     return;
   }
 
   if (!(pStat = (SRSmaStat *)tdAcquireSmaRef(smaMgmt.rsetId, pRSmaRef->refId))) {
-    smaDebug("rsma fetch task not start since rsma stat already destroyed, rsetId:%" PRIi64 " refId:%d)",
+    smaDebug("rsma fetch task not start since rsma stat already destroyed, rsetId:%d refId:%" PRIi64 ")",
              smaMgmt.rsetId, pRSmaRef->refId);  // pRSmaRef freed in taosHashRemove
     taosHashRemove(smaMgmt.refHash, &param, POINTER_BYTES);
     return;
@@ -1692,7 +1695,7 @@ static void tdRSmaFetchTrigger(void *param, void *tmrId) {
   pSma = pStat->pSma;
 
   if (!(pRSmaInfo = tdAcquireRSmaInfoBySuid(pSma, pRSmaRef->suid))) {
-    smaDebug("rsma fetch task not start since rsma info not exist, rsetId:%" PRIi64 " refId:%d)", smaMgmt.rsetId,
+    smaDebug("rsma fetch task not start since rsma info not exist, rsetId:%d refId:%" PRIi64 ")", smaMgmt.rsetId,
              pRSmaRef->refId);  // pRSmaRef freed in taosHashRemove
     tdReleaseSmaRef(smaMgmt.rsetId, pRSmaRef->refId);
     taosHashRemove(smaMgmt.refHash, &param, POINTER_BYTES);
@@ -1700,7 +1703,7 @@ static void tdRSmaFetchTrigger(void *param, void *tmrId) {
   }
 
   if (RSMA_INFO_IS_DEL(pRSmaInfo)) {
-    smaDebug("rsma fetch task not start since rsma info already deleted, rsetId:%" PRIi64 " refId:%d)", smaMgmt.rsetId,
+    smaDebug("rsma fetch task not start since rsma info already deleted, rsetId:%d refId:%" PRIi64 ")", smaMgmt.rsetId,
              pRSmaRef->refId);  // pRSmaRef freed in taosHashRemove
     tdReleaseRSmaInfo(pSma, pRSmaInfo);
     tdReleaseSmaRef(smaMgmt.rsetId, pRSmaRef->refId);
@@ -1716,7 +1719,7 @@ static void tdRSmaFetchTrigger(void *param, void *tmrId) {
     case TASK_TRIGGER_STAT_PAUSED:
     case TASK_TRIGGER_STAT_CANCELLED: {
       smaDebug("vgId:%d, rsma fetch task not start for level %" PRIi8 " since stat is %" PRIi8
-               ", rsetId rsetId:%" PRIi64 " refId:%d",
+               ", rsetId:%d refId:%" PRIi64,
                SMA_VID(pSma), pItem->level, rsmaTriggerStat, smaMgmt.rsetId, pRSmaRef->refId);
       if (rsmaTriggerStat == TASK_TRIGGER_STAT_PAUSED) {
         taosTmrReset(tdRSmaFetchTrigger, RSMA_FETCH_INTERVAL, pItem, smaMgmt.tmrHandle, &pItem->tmrId);
@@ -1844,7 +1847,7 @@ static int32_t tdRSmaBatchExec(SSma *pSma, SRSmaInfo *pInfo, STaosQall *qall, SA
     void *msg = NULL;
     taosGetQitem(qall, (void **)&msg);
     if (msg) {
-      if (taosArrayPush(pSubmitArr, &msg) < 0) {
+      if (!taosArrayPush(pSubmitArr, &msg)) {
         tdFreeRSmaSubmitItems(pSubmitArr);
         goto _err;
       }

source/dnode/vnode/src/sma/smaSnapshot.c

@@ -70,6 +70,8 @@ int32_t rsmaSnapReaderOpen(SSma* pSma, int64_t sver, int64_t ever, SRSmaSnapRead
 
   return TSDB_CODE_SUCCESS;
 _err:
+  if (pReader) rsmaSnapReaderClose(&pReader);
+  *ppReader = NULL;
   smaError("vgId:%d, vnode snapshot rsma reader open failed since %s", TD_VID(pVnode), tstrerror(code));
   return TSDB_CODE_FAILED;
 }
@@ -101,8 +103,8 @@ static int32_t rsmaQTaskInfSnapReaderOpen(SRSmaSnapReader* pReader, int64_t vers
 
   if (!taosCheckExistFile(qTaskInfoFullName)) {
     tdRSmaFSUnRef(pSma, pStat, version);
-    smaInfo("vgId:%d, vnode snapshot rsma reader for qtaskinfo version %" PRIi64 " not need as %s not exists",
-            TD_VID(pVnode), qTaskInfoFullName);
+    smaInfo("vgId:%d, vnode snapshot rsma reader for qtaskinfo version %" PRIi64 " not need as %s not exist",
+            TD_VID(pVnode), version, qTaskInfoFullName);
     return TSDB_CODE_SUCCESS;
   }
 
@@ -336,6 +338,7 @@ int32_t rsmaSnapWriterOpen(SSma* pSma, int64_t sver, int64_t ever, SRSmaSnapWrit
   tdRSmaQTaskInfoGetFullName(TD_VID(pVnode), 0, tfsGetPrimaryPath(pVnode->pTfs), qTaskInfoFullName);
   TdFilePtr qTaskF = taosCreateFile(qTaskInfoFullName, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC);
   if (!qTaskF) {
+    taosMemoryFree(qWriter);
     code = TAOS_SYSTEM_ERROR(errno);
     smaError("vgId:%d, rsma snapshot writer open %s failed since %s", TD_VID(pSma->pVnode), qTaskInfoFullName,
              tstrerror(code));
@@ -356,6 +359,7 @@ int32_t rsmaSnapWriterOpen(SSma* pSma, int64_t sver, int64_t ever, SRSmaSnapWrit
 
 _err:
   smaError("vgId:%d, rsma snapshot writer open failed since %s", TD_VID(pSma->pVnode), tstrerror(code));
+  if (pWriter) rsmaSnapWriterClose(&pWriter, 0);
   *ppWriter = NULL;
   return code;
 }
@@ -449,11 +453,11 @@ static int32_t rsmaSnapWriteQTaskInfo(SRSmaSnapWriter* pWriter, uint8_t* pData,
       code = TAOS_SYSTEM_ERROR(errno);
       goto _err;
     }
+    smaInfo("vgId:%d, vnode snapshot rsma write qtaskinfo %s succeed", SMA_VID(pWriter->pSma), qWriter->fname);
   } else {
     smaInfo("vgId:%d, vnode snapshot rsma write qtaskinfo is not needed", SMA_VID(pWriter->pSma));
   }
 
-  smaInfo("vgId:%d, vnode snapshot rsma write qtaskinfo %s succeed", SMA_VID(pWriter->pSma), qWriter->fname);
 _exit:
   return code;
 

source/dnode/vnode/src/sma/smaUtil.c

@@ -290,19 +290,19 @@ int32_t tdRemoveTFile(STFile *pTFile) {
 void *tdAcquireSmaRef(int32_t rsetId, int64_t refId) {
   void *pResult = taosAcquireRef(rsetId, refId);
   if (!pResult) {
-    smaWarn("rsma acquire ref for rsetId:%" PRIi64 " refId:%d failed since %s", rsetId, refId, terrstr());
+    smaWarn("rsma acquire ref for rsetId:%d refId:%" PRIi64 " failed since %s", rsetId, refId, terrstr());
   } else {
-    smaDebug("rsma acquire ref for rsetId:%" PRIi64 " refId:%d success", rsetId, refId);
+    smaDebug("rsma acquire ref for rsetId:%d refId:%" PRIi64 " success", rsetId, refId);
   }
   return pResult;
 }
 
 int32_t tdReleaseSmaRef(int32_t rsetId, int64_t refId) {
   if (taosReleaseRef(rsetId, refId) < 0) {
-    smaWarn("rsma release ref for rsetId:%" PRIi64 " refId:%d failed since %s", rsetId, refId, terrstr());
+    smaWarn("rsma release ref for rsetId:%d refId:%" PRIi64 " failed since %s", rsetId, refId, terrstr());
     return TSDB_CODE_FAILED;
   }
-  smaDebug("rsma release ref for rsetId:%" PRIi64 " refId:%d success", rsetId, refId);
+  smaDebug("rsma release ref for rsetId:%d refId:%" PRIi64 " success", rsetId, refId);
 
   return TSDB_CODE_SUCCESS;
 }
\ No newline at end of file

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

@@ -420,6 +420,7 @@ typedef enum {
 typedef struct {
   SFSLASTNEXTROWSTATES state;  // [input]
   STsdb               *pTsdb;  // [input]
+  STSchema            *pTSchema;// [input]
   tb_uid_t             suid;
   tb_uid_t             uid;
   int32_t              nFileSet;
@@ -455,9 +456,10 @@ static int32_t getNextRowFromFSLast(void *iter, TSDBROW **ppRow) {
       code = tsdbDataFReaderOpen(&state->pDataFReader, state->pTsdb, pFileSet);
       if (code) goto _err;
 
+      SSttBlockLoadInfo* pLoadInfo = tCreateLastBlockLoadInfo(state->pTSchema, NULL, 0);
       tMergeTreeOpen(&state->mergeTree, 1, state->pDataFReader, state->suid, state->uid,
                      &(STimeWindow){.skey = TSKEY_MIN, .ekey = TSKEY_MAX},
-                     &(SVersionRange){.minVer = 0, .maxVer = UINT64_MAX}, NULL, NULL);
+                     &(SVersionRange){.minVer = 0, .maxVer = UINT64_MAX}, pLoadInfo,true, NULL);
       bool hasVal = tMergeTreeNext(&state->mergeTree);
       if (!hasVal) {
         state->state = SFSLASTNEXTROW_FILESET;
@@ -612,7 +614,8 @@ static int32_t getNextRowFromFS(void *iter, TSDBROW **ppRow) {
         tMapDataGetItemByIdx(&state->blockMap, state->iBlock, &block, tGetDataBlk);
         /* code = tsdbReadBlockData(state->pDataFReader, &state->blockIdx, &block, &state->blockData, NULL, NULL); */
         tBlockDataReset(state->pBlockData);
-        code = tBlockDataInit(state->pBlockData, state->suid, state->uid, state->pTSchema);
+        TABLEID tid = {.suid = state->suid, .uid = state->uid};
+        code = tBlockDataInit(state->pBlockData, &tid, state->pTSchema, NULL, 0);
         if (code) goto _err;
 
         code = tsdbReadDataBlock(state->pDataFReader, &block, state->pBlockData);
@@ -891,6 +894,7 @@ static int32_t nextRowIterOpen(CacheNextRowIter *pIter, tb_uid_t uid, STsdb *pTs
   pIter->fsLastState.state = (SFSLASTNEXTROWSTATES)SFSNEXTROW_FS;
   pIter->fsLastState.pTsdb = pTsdb;
   pIter->fsLastState.aDFileSet = pIter->pReadSnap->fs.aDFileSet;
+  pIter->fsLastState.pTSchema = pTSchema;
   pIter->fsLastState.suid = suid;
   pIter->fsLastState.uid = uid;
 

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

@@ -437,7 +437,7 @@ static int32_t tsdbOpenCommitIter(SCommitter *pCommitter) {
 
         pIter->iSttBlk = 0;
         SSttBlk *pSttBlk = (SSttBlk *)taosArrayGet(pIter->aSttBlk, 0);
-        code = tsdbReadSttBlock(pCommitter->dReader.pReader, iStt, pSttBlk, &pIter->bData);
+        code = tsdbReadSttBlockEx(pCommitter->dReader.pReader, iStt, pSttBlk, &pIter->bData);
         if (code) goto _err;
 
         pIter->iRow = 0;
@@ -1049,7 +1049,7 @@ static int32_t tsdbNextCommitRow(SCommitter *pCommitter) {
         if (pIter->iSttBlk < taosArrayGetSize(pIter->aSttBlk)) {
           SSttBlk *pSttBlk = (SSttBlk *)taosArrayGet(pIter->aSttBlk, pIter->iSttBlk);
 
-          code = tsdbReadSttBlock(pCommitter->dReader.pReader, pIter->iStt, pSttBlk, &pIter->bData);
+          code = tsdbReadSttBlockEx(pCommitter->dReader.pReader, pIter->iStt, pSttBlk, &pIter->bData);
           if (code) goto _exit;
 
           pIter->iRow = 0;
@@ -1305,7 +1305,8 @@ static int32_t tsdbInitLastBlockIfNeed(SCommitter *pCommitter, TABLEID id) {
   if (!pBDatal->suid && !pBDatal->uid) {
     ASSERT(pCommitter->skmTable.suid == id.suid);
     ASSERT(pCommitter->skmTable.uid == id.uid);
-    code = tBlockDataInit(pBDatal, id.suid, id.suid ? 0 : id.uid, pCommitter->skmTable.pTSchema);
+    TABLEID tid = {.suid = id.suid, .uid = id.suid ? 0 : id.uid};
+    code = tBlockDataInit(pBDatal, &tid, pCommitter->skmTable.pTSchema, NULL, 0);
     if (code) goto _exit;
   }
 
@@ -1428,9 +1429,9 @@ static int32_t tsdbCommitFileDataImpl(SCommitter *pCommitter) {
     // impl
     code = tsdbUpdateTableSchema(pCommitter->pTsdb->pVnode->pMeta, id.suid, id.uid, &pCommitter->skmTable);
     if (code) goto _err;
-    code = tBlockDataInit(&pCommitter->dReader.bData, id.suid, id.uid, pCommitter->skmTable.pTSchema);
+    code = tBlockDataInit(&pCommitter->dReader.bData, &id, pCommitter->skmTable.pTSchema, NULL, 0);
     if (code) goto _err;
-    code = tBlockDataInit(&pCommitter->dWriter.bData, id.suid, id.uid, pCommitter->skmTable.pTSchema);
+    code = tBlockDataInit(&pCommitter->dWriter.bData, &id, pCommitter->skmTable.pTSchema, NULL, 0);
     if (code) goto _err;
 
     /* merge with data in .data file */

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

@@ -28,11 +28,10 @@ struct SLDataIter {
   uint64_t      uid;
   STimeWindow   timeWindow;
   SVersionRange verRange;
-
   SSttBlockLoadInfo* pBlockLoadInfo;
 };
 
-SSttBlockLoadInfo* tCreateLastBlockLoadInfo() {
+SSttBlockLoadInfo* tCreateLastBlockLoadInfo(STSchema* pSchema, int16_t* colList, int32_t numOfCols) {
   SSttBlockLoadInfo* pLoadInfo = taosMemoryCalloc(TSDB_DEFAULT_STT_FILE, sizeof(SSttBlockLoadInfo));
   if (pLoadInfo == NULL) {
     terrno =  TSDB_CODE_OUT_OF_MEMORY;
@@ -55,6 +54,9 @@ SSttBlockLoadInfo* tCreateLastBlockLoadInfo() {
     }
 
     pLoadInfo[i].aSttBlk = taosArrayInit(4, sizeof(SSttBlk));
+    pLoadInfo[i].pSchema = pSchema;
+    pLoadInfo[i].colIds = colList;
+    pLoadInfo[i].numOfCols = numOfCols;
   }
 
   return pLoadInfo;
@@ -111,7 +113,19 @@ static SBlockData* loadLastBlock(SLDataIter *pIter, const char* idStr) {
   pInfo->currentLoadBlockIndex ^= 1;
   if (pIter->pSttBlk != NULL) {  // current block not loaded yet
     int64_t st = taosGetTimestampUs();
-    code = tsdbReadSttBlock(pIter->pReader, pIter->iStt, pIter->pSttBlk, &pInfo->blockData[pInfo->currentLoadBlockIndex]);
+
+    SBlockData* pBlock = &pInfo->blockData[pInfo->currentLoadBlockIndex];
+
+    TABLEID id = {0};
+    if (pIter->pSttBlk->suid != 0) {
+      id.suid = pIter->pSttBlk->suid;
+    } else {
+      id.uid = pIter->uid;
+    }
+
+    tBlockDataInit(pBlock, &id, pInfo->pSchema, pInfo->colIds, pInfo->numOfCols);
+    code = tsdbReadSttBlock(pIter->pReader, pIter->iStt, pIter->pSttBlk, pBlock);
+
     double el = (taosGetTimestampUs() - st)/ 1000.0;
     pInfo->elapsedTime += el;
     pInfo->loadBlocks += 1;
@@ -460,7 +474,8 @@ static FORCE_INLINE int32_t tLDataIterCmprFn(const void *p1, const void *p2) {
 }
 
 int32_t tMergeTreeOpen(SMergeTree *pMTree, int8_t backward, SDataFReader *pFReader, uint64_t suid, uint64_t uid,
-                       STimeWindow *pTimeWindow, SVersionRange *pVerRange, void* pBlockLoadInfo, const char* idStr) {
+                       STimeWindow *pTimeWindow, SVersionRange *pVerRange, SSttBlockLoadInfo *pBlockLoadInfo,
+                       bool destroyLoadInfo, const char *idStr) {
   pMTree->backward = backward;
   pMTree->pIter = NULL;
   pMTree->pIterList = taosArrayInit(4, POINTER_BYTES);
@@ -473,21 +488,12 @@ int32_t tMergeTreeOpen(SMergeTree *pMTree, int8_t backward, SDataFReader *pFRead
   tRBTreeCreate(&pMTree->rbt, tLDataIterCmprFn);
   int32_t code = TSDB_CODE_SUCCESS;
 
-  SSttBlockLoadInfo* pLoadInfo = NULL;
-  if (pBlockLoadInfo == NULL) {
-    if (pMTree->pLoadInfo == NULL) {
-      pMTree->destroyLoadInfo = true;
-      pMTree->pLoadInfo = tCreateLastBlockLoadInfo();
-    }
-
-    pLoadInfo = pMTree->pLoadInfo;
-  } else {
-    pLoadInfo = pBlockLoadInfo;
-  }
+  pMTree->pLoadInfo = pBlockLoadInfo;
+  pMTree->destroyLoadInfo = destroyLoadInfo;
 
   for (int32_t i = 0; i < pFReader->pSet->nSttF; ++i) {  // open all last file
     struct SLDataIter* pIter = NULL;
-    code = tLDataIterOpen(&pIter, pFReader, i, pMTree->backward, suid, uid, pTimeWindow, pVerRange, &pLoadInfo[i]);
+    code = tLDataIterOpen(&pIter, pFReader, i, pMTree->backward, suid, uid, pTimeWindow, pVerRange, &pMTree->pLoadInfo[i]);
     if (code != TSDB_CODE_SUCCESS) {
       goto _end;
     }

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

@@ -85,6 +85,8 @@ _err:
 int tsdbClose(STsdb **pTsdb) {
   if (*pTsdb) {
     taosThreadRwlockDestroy(&(*pTsdb)->rwLock);
+    tsdbMemTableDestroy((*pTsdb)->mem);
+    (*pTsdb)->mem = NULL;
     tsdbFSClose(*pTsdb);
     tsdbCloseCache(*pTsdb);
     taosMemoryFreeClear(*pTsdb);

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

@@ -79,6 +79,7 @@ typedef struct SBlockLoadSuppInfo {
   SColumnDataAgg   tsColAgg;
   SColumnDataAgg** plist;
   int16_t*         colIds;    // column ids for loading file block data
+  int32_t          numOfCols;
   char**           buildBuf;  // build string tmp buffer, todo remove it later after all string format being updated.
 } SBlockLoadSuppInfo;
 
@@ -203,6 +204,7 @@ static int32_t setColumnIdSlotList(STsdbReader* pReader, SSDataBlock* pBlock) {
 
   size_t numOfCols = blockDataGetNumOfCols(pBlock);
 
+  pSupInfo->numOfCols = numOfCols;
   pSupInfo->colIds = taosMemoryMalloc(numOfCols * sizeof(int16_t));
   pSupInfo->buildBuf = taosMemoryCalloc(numOfCols, POINTER_BYTES);
   if (pSupInfo->buildBuf == NULL || pSupInfo->colIds == NULL) {
@@ -352,7 +354,8 @@ static int32_t initFilesetIterator(SFilesetIter* pIter, SArray* aDFileSet, STsdb
   tMergeTreeClose(&pLReader->mergeTree);
 
   if (pLReader->pInfo == NULL) {
-    pLReader->pInfo = tCreateLastBlockLoadInfo();
+    // here we ignore the first column, which is always be the primary timestamp column
+    pLReader->pInfo = tCreateLastBlockLoadInfo(pReader->pSchema, &pReader->suppInfo.colIds[1], pReader->suppInfo.numOfCols - 1);
     if (pLReader->pInfo == NULL) {
       tsdbDebug("init fileset iterator failed, code:%s %s", tstrerror(terrno), pReader->idStr);
       return terrno;
@@ -483,7 +486,7 @@ static int32_t tsdbReaderCreate(SVnode* pVnode, SQueryTableDataCond* pCond, STsd
   pReader->pTsdb = getTsdbByRetentions(pVnode, pCond->twindows.skey, pVnode->config.tsdbCfg.retentions, idstr, &level);
   pReader->suid = pCond->suid;
   pReader->order = pCond->order;
-  pReader->capacity = 4096;
+  pReader->capacity = capacity;
   pReader->idStr = (idstr != NULL) ? strdup(idstr) : NULL;
   pReader->verRange = getQueryVerRange(pVnode, pCond, level);
   pReader->type = pCond->type;
@@ -838,14 +841,18 @@ static int32_t copyBlockDataToSDataBlock(STsdbReader* pReader, STableBlockScanIn
   bool    asc = ASCENDING_TRAVERSE(pReader->order);
   int32_t step = asc ? 1 : -1;
 
-  if (asc && pReader->window.skey <= pBlock->minKey.ts) {
-    pDumpInfo->rowIndex = 0;
-  } else if (!asc && pReader->window.ekey >= pBlock->maxKey.ts) {
-    pDumpInfo->rowIndex = pBlock->nRow - 1;
-  } else {
-    int32_t pos = asc ? pBlock->nRow - 1 : 0;
-    int32_t order = (pReader->order == TSDB_ORDER_ASC) ? TSDB_ORDER_DESC : TSDB_ORDER_ASC;
-    pDumpInfo->rowIndex = doBinarySearchKey(pBlockData->aTSKEY, pBlock->nRow, pos, pReader->window.skey, order);
+
+  if ((pDumpInfo->rowIndex == 0 && asc) || (pDumpInfo->rowIndex == pBlock->nRow - 1 && (!asc))) {
+    if (asc && pReader->window.skey <= pBlock->minKey.ts) {
+      //pDumpInfo->rowIndex = 0;
+    } else 
+    if (!asc && pReader->window.ekey >= pBlock->maxKey.ts) {
+      //pDumpInfo->rowIndex = pBlock->nRow - 1;
+    } else {
+      int32_t pos = asc ? pBlock->nRow - 1 : 0;
+      int32_t order = (pReader->order == TSDB_ORDER_ASC) ? TSDB_ORDER_DESC : TSDB_ORDER_ASC;
+      pDumpInfo->rowIndex = doBinarySearchKey(pBlockData->aTSKEY, pBlock->nRow, pos, pReader->window.skey, order);
+    }
   }
 
   // time window check
@@ -929,8 +936,8 @@ static int32_t copyBlockDataToSDataBlock(STsdbReader* pReader, STableBlockScanIn
   pDumpInfo->rowIndex += step * remain;
 
   if (pDumpInfo->rowIndex >= 0 && pDumpInfo->rowIndex < pBlock->nRow) {
-    int64_t ts = pBlockData->aTSKEY[pDumpInfo->rowIndex];
-    setBlockAllDumped(pDumpInfo, ts, pReader->order);
+//    int64_t ts = pBlockData->aTSKEY[pDumpInfo->rowIndex];
+//    setBlockAllDumped(pDumpInfo, ts, pReader->order);
   } else {
     int64_t k = asc ? pBlock->maxKey.ts : pBlock->minKey.ts;
     setBlockAllDumped(pDumpInfo, k, pReader->order);
@@ -948,15 +955,22 @@ static int32_t copyBlockDataToSDataBlock(STsdbReader* pReader, STableBlockScanIn
   return TSDB_CODE_SUCCESS;
 }
 
-static int32_t doLoadFileBlockData(STsdbReader* pReader, SDataBlockIter* pBlockIter, SBlockData* pBlockData) {
+static int32_t doLoadFileBlockData(STsdbReader* pReader, SDataBlockIter* pBlockIter, SBlockData* pBlockData, uint64_t uid) {
   int64_t st = taosGetTimestampUs();
 
+  tBlockDataReset(pBlockData);
+  TABLEID tid = {.suid = pReader->suid, .uid = uid};
+  int32_t code = tBlockDataInit(pBlockData, &tid, pReader->pSchema, &pReader->suppInfo.colIds[1], pReader->suppInfo.numOfCols-1);
+  if (code != TSDB_CODE_SUCCESS) {
+    return code;
+  }
+
   SFileDataBlockInfo* pBlockInfo = getCurrentBlockInfo(pBlockIter);
   SFileBlockDumpInfo* pDumpInfo = &pReader->status.fBlockDumpInfo;
   ASSERT(pBlockInfo != NULL);
 
   SDataBlk* pBlock = getCurrentBlock(pBlockIter);
-  int32_t   code = tsdbReadDataBlock(pReader->pFileReader, pBlock, pBlockData);
+  code = tsdbReadDataBlock(pReader->pFileReader, pBlock, pBlockData);
   if (code != TSDB_CODE_SUCCESS) {
     tsdbError("%p error occurs in loading file block, global index:%d, table index:%d, brange:%" PRId64 "-%" PRId64
               ", rows:%d, code:%s %s",
@@ -1741,7 +1755,7 @@ static int32_t doMergeMultiLevelRows(STsdbReader* pReader, STableBlockScanInfo*
     tsLast = getCurrentKeyInLastBlock(pLastBlockReader);
   }
 
-  int64_t key = hasDataInFileBlock(pBlockData, pDumpInfo)? pBlockData->aTSKEY[pDumpInfo->rowIndex]:INT64_MIN;
+  int64_t key = hasDataInFileBlock(pBlockData, pDumpInfo) ? pBlockData->aTSKEY[pDumpInfo->rowIndex] : INT64_MIN;
 
   TSDBKEY k = TSDBROW_KEY(pRow);
   TSDBKEY ik = TSDBROW_KEY(piRow);
@@ -1995,7 +2009,7 @@ static bool initLastBlockReader(SLastBlockReader* pLBlockReader, STableBlockScan
 
   int32_t code =
       tMergeTreeOpen(&pLBlockReader->mergeTree, (pLBlockReader->order == TSDB_ORDER_DESC), pReader->pFileReader,
-                     pReader->suid, pScanInfo->uid, &w, &pLBlockReader->verRange, pLBlockReader->pInfo, pReader->idStr);
+                     pReader->suid, pScanInfo->uid, &w, &pLBlockReader->verRange, pLBlockReader->pInfo, false, pReader->idStr);
   if (code != TSDB_CODE_SUCCESS) {
     return false;
   }
@@ -2009,12 +2023,12 @@ static int64_t getCurrentKeyInLastBlock(SLastBlockReader* pLastBlockReader) {
 }
 
 static bool hasDataInLastBlock(SLastBlockReader* pLastBlockReader) { return pLastBlockReader->mergeTree.pIter != NULL; }
-bool hasDataInFileBlock(const SBlockData* pBlockData, const SFileBlockDumpInfo* pDumpInfo) {
-  if (pBlockData->nRow > 0) {
-    ASSERT(pBlockData->nRow == pDumpInfo->totalRows);
+bool        hasDataInFileBlock(const SBlockData* pBlockData, const SFileBlockDumpInfo* pDumpInfo) {
+         if (pBlockData->nRow > 0) {
+           ASSERT(pBlockData->nRow == pDumpInfo->totalRows);
   }
 
-  return pBlockData->nRow > 0 && (!pDumpInfo->allDumped);
+         return pBlockData->nRow > 0 && (!pDumpInfo->allDumped);
 }
 
 int32_t mergeRowsInFileBlocks(SBlockData* pBlockData, STableBlockScanInfo* pBlockScanInfo, int64_t key,
@@ -2451,13 +2465,7 @@ static int32_t doBuildDataBlock(STsdbReader* pReader) {
     ASSERT(pBlockIter->numOfBlocks == 0);
     code = buildComposedDataBlock(pReader);
   } else if (fileBlockShouldLoad(pReader, pBlockInfo, pBlock, pScanInfo, keyInBuf, pLastBlockReader)) {
-    tBlockDataReset(&pStatus->fileBlockData);
-    code = tBlockDataInit(&pStatus->fileBlockData, pReader->suid, pScanInfo->uid, pReader->pSchema);
-    if (code != TSDB_CODE_SUCCESS) {
-      return code;
-    }
-
-    code = doLoadFileBlockData(pReader, pBlockIter, &pStatus->fileBlockData);
+    code = doLoadFileBlockData(pReader, pBlockIter, &pStatus->fileBlockData, pScanInfo->uid);
     if (code != TSDB_CODE_SUCCESS) {
       return code;
     }
@@ -2931,13 +2939,7 @@ static int32_t checkForNeighborFileBlock(STsdbReader* pReader, STableBlockScanIn
     setFileBlockActiveInBlockIter(pBlockIter, neighborIndex, step);
 
     // 3. load the neighbor block, and set it to be the currently accessed file data block
-    tBlockDataReset(&pStatus->fileBlockData);
-    int32_t code = tBlockDataInit(&pStatus->fileBlockData, pReader->suid, pFBlock->uid, pReader->pSchema);
-    if (code != TSDB_CODE_SUCCESS) {
-      return code;
-    }
-
-    code = doLoadFileBlockData(pReader, pBlockIter, &pStatus->fileBlockData);
+    int32_t code = doLoadFileBlockData(pReader, pBlockIter, &pStatus->fileBlockData, pFBlock->uid);
     if (code != TSDB_CODE_SUCCESS) {
       return code;
     }
@@ -3404,10 +3406,14 @@ int32_t tsdbReaderOpen(SVnode* pVnode, SQueryTableDataCond* pCond, SArray* pTabl
     // we need only one row
     pPrevReader->capacity = 1;
     pPrevReader->status.pTableMap = pReader->status.pTableMap;
+    pPrevReader->pSchema = pReader->pSchema;
+    pPrevReader->pMemSchema = pReader->pMemSchema;
     pPrevReader->pReadSnap = pReader->pReadSnap;
 
     pNextReader->capacity = 1;
     pNextReader->status.pTableMap = pReader->status.pTableMap;
+    pNextReader->pSchema = pReader->pSchema;
+    pNextReader->pMemSchema = pReader->pMemSchema;
     pNextReader->pReadSnap = pReader->pReadSnap;
 
     code = doOpenReaderImpl(pPrevReader);
@@ -3441,11 +3447,19 @@ void tsdbReaderClose(STsdbReader* pReader) {
 
   {
     if (pReader->innerReader[0] != NULL) {
-      pReader->innerReader[0]->status.pTableMap = NULL;
-      pReader->innerReader[0]->pReadSnap = NULL;
+      STsdbReader* p = pReader->innerReader[0];
+
+      p->status.pTableMap = NULL;
+      p->pReadSnap = NULL;
+      p->pSchema = NULL;
+      p->pMemSchema = NULL;
+
+      p = pReader->innerReader[1];
 
-      pReader->innerReader[1]->status.pTableMap = NULL;
-      pReader->innerReader[1]->pReadSnap = NULL;
+      p->status.pTableMap = NULL;
+      p->pReadSnap = NULL;
+      p->pSchema = NULL;
+      p->pMemSchema = NULL;
 
       tsdbReaderClose(pReader->innerReader[0]);
       tsdbReaderClose(pReader->innerReader[1]);
@@ -3683,14 +3697,7 @@ static SArray* doRetrieveDataBlock(STsdbReader* pReader) {
   SFileDataBlockInfo*  pFBlock = getCurrentBlockInfo(&pStatus->blockIter);
   STableBlockScanInfo* pBlockScanInfo = taosHashGet(pStatus->pTableMap, &pFBlock->uid, sizeof(pFBlock->uid));
 
-  tBlockDataReset(&pStatus->fileBlockData);
-  int32_t code = tBlockDataInit(&pStatus->fileBlockData, pReader->suid, pBlockScanInfo->uid, pReader->pSchema);
-  if (code != TSDB_CODE_SUCCESS) {
-    terrno = code;
-    return NULL;
-  }
-
-  code = doLoadFileBlockData(pReader, &pStatus->blockIter, &pStatus->fileBlockData);
+  int32_t code = doLoadFileBlockData(pReader, &pStatus->blockIter, &pStatus->fileBlockData, pBlockScanInfo->uid);
   if (code != TSDB_CODE_SUCCESS) {
     tBlockDataDestroy(&pStatus->fileBlockData, 1);
     terrno = code;

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

@@ -926,12 +926,13 @@ _err:
   return code;
 }
 
-static int32_t tsdbReadBlockDataImpl(SDataFReader *pReader, SBlockInfo *pBlkInfo, SBlockData *pBlockData) {
+static int32_t tsdbReadBlockDataImpl(SDataFReader *pReader, SBlockInfo *pBlkInfo, SBlockData *pBlockData,
+                                     int32_t iStt) {
   int32_t code = 0;
 
   tBlockDataClear(pBlockData);
 
-  STsdbFD *pFD = pReader->pDataFD;
+  STsdbFD *pFD = (iStt < 0) ? pReader->pDataFD : pReader->aSttFD[iStt];
 
   // uid + version + tskey
   code = tRealloc(&pReader->aBuf[0], pBlkInfo->szKey);
@@ -1070,9 +1071,12 @@ _err:
 int32_t tsdbReadDataBlock(SDataFReader *pReader, SDataBlk *pDataBlk, SBlockData *pBlockData) {
   int32_t code = 0;
 
-  code = tsdbReadBlockDataImpl(pReader, &pDataBlk->aSubBlock[0], pBlockData);
+  code = tsdbReadBlockDataImpl(pReader, &pDataBlk->aSubBlock[0], pBlockData, -1);
   if (code) goto _err;
 
+  ASSERT(pDataBlk->nSubBlock == 1);
+
+#if 0
   if (pDataBlk->nSubBlock > 1) {
     SBlockData bData1;
     SBlockData bData2;
@@ -1113,6 +1117,7 @@ int32_t tsdbReadDataBlock(SDataFReader *pReader, SDataBlk *pDataBlk, SBlockData
     tBlockDataDestroy(&bData1, 1);
     tBlockDataDestroy(&bData2, 1);
   }
+#endif
 
   return code;
 
@@ -1123,23 +1128,38 @@ _err:
 
 int32_t tsdbReadSttBlock(SDataFReader *pReader, int32_t iStt, SSttBlk *pSttBlk, SBlockData *pBlockData) {
   int32_t code = 0;
+  int32_t lino = 0;
+
+  code = tsdbReadBlockDataImpl(pReader, &pSttBlk->bInfo, pBlockData, iStt);
+  TSDB_CHECK_CODE(code, lino, _exit);
+
+_exit:
+  if (code) {
+    tsdbError("vgId:%d %s failed at %d since %s", TD_VID(pReader->pTsdb->pVnode), __func__, lino, tstrerror(code));
+  }
+  return code;
+}
+
+int32_t tsdbReadSttBlockEx(SDataFReader *pReader, int32_t iStt, SSttBlk *pSttBlk, SBlockData *pBlockData) {
+  int32_t code = 0;
+  int32_t lino = 0;
 
   // alloc
   code = tRealloc(&pReader->aBuf[0], pSttBlk->bInfo.szBlock);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // read
   code = tsdbReadFile(pReader->aSttFD[iStt], pSttBlk->bInfo.offset, pReader->aBuf[0], pSttBlk->bInfo.szBlock);
-  if (code) goto _err;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
   // decmpr
   code = tDecmprBlockData(pReader->aBuf[0], pSttBlk->bInfo.szBlock, pBlockData, &pReader->aBuf[1]);
-  if (code) goto _err;
-
-  return code;
+  TSDB_CHECK_CODE(code, lino, _exit);
 
-_err:
-  tsdbError("vgId:%d tsdb read stt block failed since %s", TD_VID(pReader->pTsdb->pVnode), tstrerror(code));
+_exit:
+  if (code) {
+    tsdbError("vgId:%d %s failed at %d since %s", TD_VID(pReader->pTsdb->pVnode), __func__, lino, tstrerror(code));
+  }
   return code;
 }
 

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

@@ -140,7 +140,7 @@ static int32_t tsdbSnapReadOpenFile(STsdbSnapReader* pReader) {
       if (pSttBlk->minVer > pReader->ever) continue;
       if (pSttBlk->maxVer < pReader->sver) continue;
 
-      code = tsdbReadSttBlock(pReader->pDataFReader, iStt, pSttBlk, &pIter->bData);
+      code = tsdbReadSttBlockEx(pReader->pDataFReader, iStt, pSttBlk, &pIter->bData);
       if (code) goto _err;
 
       for (pIter->iRow = 0; pIter->iRow < pIter->bData.nRow; pIter->iRow++) {
@@ -223,7 +223,7 @@ static int32_t tsdbSnapNextRow(STsdbSnapReader* pReader) {
 
           if (pSttBlk->minVer > pReader->ever || pSttBlk->maxVer < pReader->sver) continue;
 
-          code = tsdbReadSttBlock(pReader->pDataFReader, pIter->iStt, pSttBlk, &pIter->bData);
+          code = tsdbReadSttBlockEx(pReader->pDataFReader, pIter->iStt, pSttBlk, &pIter->bData);
           if (code) goto _err;
 
           pIter->iRow = -1;
@@ -319,7 +319,7 @@ static int32_t tsdbSnapReadData(STsdbSnapReader* pReader, uint8_t** ppData) {
     code = tsdbUpdateTableSchema(pTsdb->pVnode->pMeta, id.suid, id.uid, &pReader->skmTable);
     if (code) goto _err;
 
-    code = tBlockDataInit(pBlockData, id.suid, id.uid, pReader->skmTable.pTSchema);
+    code = tBlockDataInit(pBlockData, &id, pReader->skmTable.pTSchema, NULL, 0);
     if (code) goto _err;
 
     while (pRowInfo->suid == id.suid && pRowInfo->uid == id.uid) {
@@ -715,7 +715,7 @@ static int32_t tsdbSnapWriteTableDataStart(STsdbSnapWriter* pWriter, TABLEID* pI
   if (code) goto _err;
 
   tMapDataReset(&pWriter->dWriter.mDataBlk);
-  code = tBlockDataInit(&pWriter->dWriter.bData, pId->suid, pId->uid, pWriter->skmTable.pTSchema);
+  code = tBlockDataInit(&pWriter->dWriter.bData, pId, pWriter->skmTable.pTSchema, NULL, 0);
   if (code) goto _err;
 
   return code;
@@ -1000,7 +1000,8 @@ static int32_t tsdbSnapWriteToSttFile(STsdbSnapWriter* pWriter, int32_t iRow) {
     code = tsdbUpdateTableSchema(pWriter->pTsdb->pVnode->pMeta, pWriter->id.suid, pWriter->id.uid, &pWriter->skmTable);
     if (code) goto _err;
 
-    code = tBlockDataInit(pBData, pWriter->id.suid, pWriter->id.suid ? 0 : pWriter->id.uid, pWriter->skmTable.pTSchema);
+    TABLEID tid = {.suid = pWriter->id.suid, .uid = pWriter->id.suid ? 0 : pWriter->id.uid};
+    code = tBlockDataInit(pBData, &tid, pWriter->skmTable.pTSchema, NULL, 0);
     if (code) goto _err;
   }
 

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

@@ -948,24 +948,47 @@ void tBlockDataDestroy(SBlockData *pBlockData, int8_t deepClear) {
   pBlockData->aColData = NULL;
 }
 
-int32_t tBlockDataInit(SBlockData *pBlockData, int64_t suid, int64_t uid, STSchema *pTSchema) {
+int32_t tBlockDataInit(SBlockData *pBlockData, TABLEID *pId, STSchema *pTSchema, int16_t *aCid, int32_t nCid) {
   int32_t code = 0;
 
-  ASSERT(suid || uid);
+  ASSERT(pId->suid || pId->uid);
 
-  pBlockData->suid = suid;
-  pBlockData->uid = uid;
+  pBlockData->suid = pId->suid;
+  pBlockData->uid = pId->uid;
   pBlockData->nRow = 0;
 
   taosArrayClear(pBlockData->aIdx);
-  for (int32_t iColumn = 1; iColumn < pTSchema->numOfCols; iColumn++) {
+  if (aCid) {
+    int32_t   iColumn = 1;
     STColumn *pTColumn = &pTSchema->columns[iColumn];
+    for (int32_t iCid = 0; iCid < nCid; iCid++) {
+      while (pTColumn && pTColumn->colId < aCid[iCid]) {
+        iColumn++;
+        pTColumn = (iColumn < pTSchema->numOfCols) ? &pTSchema->columns[iColumn] : NULL;
+      }
 
-    SColData *pColData;
-    code = tBlockDataAddColData(pBlockData, iColumn - 1, &pColData);
-    if (code) goto _exit;
+      if (pTColumn == NULL) {
+        break;
+      } else if (pTColumn->colId == aCid[iCid]) {
+        SColData *pColData;
+        code = tBlockDataAddColData(pBlockData, taosArrayGetSize(pBlockData->aIdx), &pColData);
+        if (code) goto _exit;
+        tColDataInit(pColData, pTColumn->colId, pTColumn->type, (pTColumn->flags & COL_SMA_ON) ? 1 : 0);
 
-    tColDataInit(pColData, pTColumn->colId, pTColumn->type, (pTColumn->flags & COL_SMA_ON) ? 1 : 0);
+        iColumn++;
+        pTColumn = (iColumn < pTSchema->numOfCols) ? &pTSchema->columns[iColumn] : NULL;
+      }
+    }
+  } else {
+    for (int32_t iColumn = 1; iColumn < pTSchema->numOfCols; iColumn++) {
+      STColumn *pTColumn = &pTSchema->columns[iColumn];
+
+      SColData *pColData;
+      code = tBlockDataAddColData(pBlockData, iColumn - 1, &pColData);
+      if (code) goto _exit;
+
+      tColDataInit(pColData, pTColumn->colId, pTColumn->type, (pTColumn->flags & COL_SMA_ON) ? 1 : 0);
+    }
   }
 
 _exit:

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

@@ -53,6 +53,10 @@ int vnodeCloseBufPool(SVnode *pVnode) {
     vnodeBufPoolDestroy(pPool);
   }
 
+  if (pVnode->inUse) {
+    vnodeBufPoolDestroy(pVnode->inUse);
+    pVnode->inUse = NULL;
+  }
   vDebug("vgId:%d, vnode buffer pool is closed", TD_VID(pVnode));
 
   return 0;
@@ -177,4 +181,4 @@ void vnodeBufPoolUnRef(SVBufPool *pPool) {
 
     taosThreadMutexUnlock(&pVnode->mutex);
   }
-}
+}
\ No newline at end of file

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

@@ -161,7 +161,6 @@ SVnode *vnodeOpen(const char *path, STfs *pTfs, SMsgCb msgCb) {
   // open sync
   if (vnodeSyncOpen(pVnode, dir)) {
     vError("vgId:%d, failed to open sync since %s", TD_VID(pVnode), tstrerror(terrno));
-    terrno = TSDB_CODE_OUT_OF_MEMORY;
     goto _err;
   }
 
@@ -174,6 +173,7 @@ _err:
   if (pVnode->pTsdb) tsdbClose(&pVnode->pTsdb);
   if (pVnode->pSma) smaClose(pVnode->pSma);
   if (pVnode->pMeta) metaClose(pVnode->pMeta);
+  if (pVnode->pPool) vnodeCloseBufPool(pVnode);
 
   tsem_destroy(&(pVnode->canCommit));
   taosMemoryFree(pVnode);

source/libs/executor/inc/executorimpl.h

@@ -34,6 +34,7 @@ extern "C" {
 #include "scalar.h"
 #include "taosdef.h"
 #include "tarray.h"
+#include "tfill.h"
 #include "thash.h"
 #include "tlockfree.h"
 #include "tmsg.h"
@@ -798,6 +799,22 @@ typedef struct SStreamPartitionOperatorInfo {
   SSDataBlock*          pDelRes;
 } SStreamPartitionOperatorInfo;
 
+typedef struct SStreamFillOperatorInfo {
+  SStreamFillSupporter* pFillSup;
+  SSDataBlock*      pRes;
+  SSDataBlock*      pSrcBlock;
+  int32_t           srcRowIndex;
+  SSDataBlock*      pPrevSrcBlock;
+  SSDataBlock*      pSrcDelBlock;
+  int32_t           srcDelRowIndex;
+  SSDataBlock*      pDelRes;
+  SNode*            pCondition;
+  SArray*           pColMatchColInfo;
+  int32_t           primaryTsCol;
+  int32_t           primarySrcSlotId;
+  SStreamFillInfo*  pFillInfo;
+} SStreamFillOperatorInfo;
+
 typedef struct STimeSliceOperatorInfo {
   SSDataBlock*         pRes;
   STimeWindow          win;
@@ -1006,6 +1023,8 @@ SOperatorInfo* createStreamIntervalOperatorInfo(SOperatorInfo* downstream,
 
 SOperatorInfo* createStreamStateAggOperatorInfo(SOperatorInfo* downstream, SPhysiNode* pPhyNode,
                                                 SExecTaskInfo* pTaskInfo);
+SOperatorInfo* createStreamFillOperatorInfo(SOperatorInfo* downstream, SStreamFillPhysiNode* pPhyFillNode,
+                                            SExecTaskInfo* pTaskInfo);
 
 int32_t projectApplyFunctions(SExprInfo* pExpr, SSDataBlock* pResult, SSDataBlock* pSrcBlock, SqlFunctionCtx* pCtx,
                               int32_t numOfOutput, SArray* pPseudoList);
@@ -1094,6 +1113,7 @@ int32_t setOutputBuf(STimeWindow* win, SResultRow** pResult, int64_t tableGroupI
                            SExecTaskInfo* pTaskInfo);
 int32_t releaseOutputBuf(SExecTaskInfo* pTaskInfo, SWinKey* pKey, SResultRow* pResult);
 int32_t saveOutputBuf(SExecTaskInfo* pTaskInfo, SWinKey* pKey, SResultRow* pResult, int32_t resSize);
+void    getNextIntervalWindow(SInterval* pInterval, STimeWindow* tw, int32_t order);
 
 #ifdef __cplusplus
 }

source/libs/executor/inc/tfill.h

@@ -23,12 +23,13 @@ extern "C" {
 #include "os.h"
 #include "taosdef.h"
 #include "tcommon.h"
+#include "tsimplehash.h"
 
 struct SSDataBlock;
 
 typedef struct SFillColInfo {
-  SExprInfo *pExpr;
-  bool       notFillCol;      // denote if this column needs fill operation
+  SExprInfo* pExpr;
+  bool       notFillCol;  // denote if this column needs fill operation
   SVariant   fillVal;
 } SFillColInfo;
 
@@ -51,46 +52,96 @@ typedef struct {
 } SRowVal;
 
 typedef struct SFillInfo {
-  TSKEY     start;                // start timestamp
-  TSKEY     end;                  // endKey for fill
-  TSKEY     currentKey;           // current active timestamp, the value may be changed during the fill procedure.
-  int32_t   tsSlotId;             // primary time stamp slot id
-  int32_t   srcTsSlotId;          // timestamp column id in the source data block.
-  int32_t   order;                // order [TSDB_ORDER_ASC|TSDB_ORDER_DESC]
-  int32_t   type;                 // fill type
-  int32_t   numOfRows;            // number of rows in the input data block
-  int32_t   index;                // active row index
-  int32_t   numOfTotal;           // number of filled rows in one round
-  int32_t   numOfCurrent;         // number of filled rows in current results
-  int32_t   numOfCols;            // number of columns, including the tags columns
-  SInterval interval;
-  SRowVal   prev;
-  SRowVal   next;
-  SSDataBlock *pSrcBlock;
-  int32_t   alloc;                // data buffer size in rows
-
-  SFillColInfo* pFillCol;         // column info for fill operations
-  SFillTagColInfo* pTags;         // tags value for filling gap
-  const char* id;
+  TSKEY        start;         // start timestamp
+  TSKEY        end;           // endKey for fill
+  TSKEY        currentKey;    // current active timestamp, the value may be changed during the fill procedure.
+  int32_t      tsSlotId;      // primary time stamp slot id
+  int32_t      srcTsSlotId;   // timestamp column id in the source data block.
+  int32_t      order;         // order [TSDB_ORDER_ASC|TSDB_ORDER_DESC]
+  int32_t      type;          // fill type
+  int32_t      numOfRows;     // number of rows in the input data block
+  int32_t      index;         // active row index
+  int32_t      numOfTotal;    // number of filled rows in one round
+  int32_t      numOfCurrent;  // number of filled rows in current results
+  int32_t      numOfCols;     // number of columns, including the tags columns
+  SInterval    interval;
+  SRowVal      prev;
+  SRowVal      next;
+  SSDataBlock* pSrcBlock;
+  int32_t      alloc;  // data buffer size in rows
+
+  SFillColInfo*    pFillCol;  // column info for fill operations
+  SFillTagColInfo* pTags;     // tags value for filling gap
+  const char*      id;
 } SFillInfo;
 
-int64_t getNumOfResultsAfterFillGap(SFillInfo* pFillInfo, int64_t ekey, int32_t maxNumOfRows);
+typedef struct SResultCellData {
+  bool    isNull;
+  int8_t  type;
+  int32_t bytes;
+  char    pData[];
+} SResultCellData;
+
+typedef struct SResultRowData {
+  TSKEY            key;
+  SResultCellData* pRowVal;
+} SResultRowData;
+
+typedef struct SStreamFillLinearInfo {
+  TSKEY   nextEnd;
+  SArray* pDeltaVal;      // double. value for Fill(linear).
+  SArray* pNextDeltaVal;  // double. value for Fill(linear).
+  int64_t winIndex;
+  bool    hasNext;
+} SStreamFillLinearInfo;
+
+typedef struct SStreamFillInfo {
+  TSKEY                  start;    // startKey for fill
+  TSKEY                  end;      // endKey for fill
+  TSKEY                  current;  // current Key for fill
+  TSKEY                  preRowKey;
+  TSKEY                  nextRowKey;
+  SResultRowData*        pResRow;
+  SStreamFillLinearInfo* pLinearInfo;
+  bool                   needFill;
+  int32_t                type;  // fill type
+  int32_t                pos;
+  SArray*                delRanges;
+  int32_t                delIndex;
+} SStreamFillInfo;
+
+typedef struct SStreamFillSupporter {
+  int32_t        type;  // fill type
+  SInterval      interval;
+  SResultRowData prev;
+  SResultRowData cur;
+  SResultRowData next;
+  SResultRowData nextNext;
+  SFillColInfo*  pAllColInfo;   // fill exprs and not fill exprs
+  int32_t        numOfAllCols;  // number of all exprs, including the tags columns
+  int32_t        numOfFillCols;
+  int32_t        numOfNotFillCols;
+  int32_t        rowSize;
+  SSHashObj*     pResMap;
+  bool           hasDelete;
+} SStreamFillSupporter;
 
+int64_t getNumOfResultsAfterFillGap(SFillInfo* pFillInfo, int64_t ekey, int32_t maxNumOfRows);
 
-void taosFillSetStartInfo(struct SFillInfo* pFillInfo, int32_t numOfRows, TSKEY endKey);
-void taosResetFillInfo(struct SFillInfo* pFillInfo, TSKEY startTimestamp);
-void taosFillSetInputDataBlock(struct SFillInfo* pFillInfo, const struct SSDataBlock* pInput);
-struct SFillColInfo* createFillColInfo(SExprInfo* pExpr, int32_t numOfFillExpr, SExprInfo* pNotFillExpr, int32_t numOfNotFillCols, const struct SNodeListNode* val);
-bool taosFillHasMoreResults(struct SFillInfo* pFillInfo);
+void                 taosFillSetStartInfo(struct SFillInfo* pFillInfo, int32_t numOfRows, TSKEY endKey);
+void                 taosResetFillInfo(struct SFillInfo* pFillInfo, TSKEY startTimestamp);
+void                 taosFillSetInputDataBlock(struct SFillInfo* pFillInfo, const struct SSDataBlock* pInput);
+struct SFillColInfo* createFillColInfo(SExprInfo* pExpr, int32_t numOfFillExpr, SExprInfo* pNotFillExpr,
+                                       int32_t numOfNotFillCols, const struct SNodeListNode* val);
+bool                 taosFillHasMoreResults(struct SFillInfo* pFillInfo);
 
 SFillInfo* taosCreateFillInfo(TSKEY skey, int32_t numOfFillCols, int32_t numOfNotFillCols, int32_t capacity,
                               SInterval* pInterval, int32_t fillType, struct SFillColInfo* pCol, int32_t slotId,
                               int32_t order, const char* id);
 
-void* taosDestroyFillInfo(struct SFillInfo *pFillInfo);
+void*   taosDestroyFillInfo(struct SFillInfo* pFillInfo);
 int64_t taosFillResultDataBlock(struct SFillInfo* pFillInfo, SSDataBlock* p, int32_t capacity);
-int64_t getFillInfoStart(struct SFillInfo *pFillInfo);
-
+int64_t getFillInfoStart(struct SFillInfo* pFillInfo);
 
 #ifdef __cplusplus
 }

source/libs/executor/src/executor.c

@@ -479,7 +479,7 @@ static void freeBlock(void* param) {
   blockDataDestroy(pBlock);
 }
 
-int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, SLocalFetch* pLocal) {
+int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, bool* hasMore, SLocalFetch* pLocal) {
   SExecTaskInfo* pTaskInfo = (SExecTaskInfo*)tinfo;
   int64_t        threadId = taosGetSelfPthreadId();
 
@@ -536,6 +536,7 @@ int32_t qExecTaskOpt(qTaskInfo_t tinfo, SArray* pResList, uint64_t* useconds, SL
     }
   }
 
+  *hasMore = (pRes != NULL);
   uint64_t el = (taosGetTimestampUs() - st);
 
   pTaskInfo->cost.elapsedTime += el;

source/libs/executor/src/executorimpl.c

@@ -3742,6 +3742,8 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
     pOptr = createMergeJoinOperatorInfo(ops, size, (SSortMergeJoinPhysiNode*)pPhyNode, pTaskInfo);
   } else if (QUERY_NODE_PHYSICAL_PLAN_FILL == type) {
     pOptr = createFillOperatorInfo(ops[0], (SFillPhysiNode*)pPhyNode, pTaskInfo);
+  } else if (QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL == type) {
+    pOptr = createStreamFillOperatorInfo(ops[0], (SStreamFillPhysiNode*)pPhyNode, pTaskInfo);
   } else if (QUERY_NODE_PHYSICAL_PLAN_INDEF_ROWS_FUNC == type) {
     pOptr = createIndefinitOutputOperatorInfo(ops[0], pPhyNode, pTaskInfo);
   } else if (QUERY_NODE_PHYSICAL_PLAN_INTERP_FUNC == type) {

source/libs/executor/src/scanoperator.c

@@ -1022,11 +1022,7 @@ static uint64_t getGroupIdByCol(SStreamScanInfo* pInfo, uint64_t uid, TSKEY ts,
   return calGroupIdByData(&pInfo->partitionSup, pInfo->pPartScalarSup, pPreRes, 0);
 }
 
-static uint64_t getGroupIdByData(SStreamScanInfo* pInfo, uint64_t uid, TSKEY ts, int64_t maxVersion) {
-  if (pInfo->partitionSup.needCalc) {
-    return getGroupIdByCol(pInfo, uid, ts, maxVersion);
-  }
-
+static uint64_t getGroupIdByUid(SStreamScanInfo* pInfo, uint64_t uid) {
   SHashObj* map = pInfo->pTableScanOp->pTaskInfo->tableqinfoList.map;
   uint64_t* groupId = taosHashGet(map, &uid, sizeof(int64_t));
   if (groupId) {
@@ -1035,6 +1031,14 @@ static uint64_t getGroupIdByData(SStreamScanInfo* pInfo, uint64_t uid, TSKEY ts,
   return 0;
 }
 
+static uint64_t getGroupIdByData(SStreamScanInfo* pInfo, uint64_t uid, TSKEY ts, int64_t maxVersion) {
+  if (pInfo->partitionSup.needCalc) {
+    return getGroupIdByCol(pInfo, uid, ts, maxVersion);
+  }
+
+  return getGroupIdByUid(pInfo, uid);
+}
+
 static bool prepareRangeScan(SStreamScanInfo* pInfo, SSDataBlock* pBlock, int32_t* pRowIndex) {
   if ((*pRowIndex) == pBlock->info.rows) {
     return false;
@@ -1081,26 +1085,32 @@ static bool prepareRangeScan(SStreamScanInfo* pInfo, SSDataBlock* pBlock, int32_
   return true;
 }
 
-static STimeWindow getSlidingWindow(TSKEY* startTsCol, TSKEY* endTsCol, SInterval* pInterval,
+static STimeWindow getSlidingWindow(TSKEY* startTsCol, TSKEY* endTsCol, uint64_t* gpIdCol, SInterval* pInterval,
                                     SDataBlockInfo* pDataBlockInfo, int32_t* pRowIndex, bool hasGroup) {
   SResultRowInfo dumyInfo;
   dumyInfo.cur.pageId = -1;
   STimeWindow win = getActiveTimeWindow(NULL, &dumyInfo, startTsCol[*pRowIndex], pInterval, TSDB_ORDER_ASC);
   STimeWindow endWin = win;
   STimeWindow preWin = win;
+  uint64_t    groupId = gpIdCol[*pRowIndex];
   while (1) {
     if (hasGroup) {
       (*pRowIndex) += 1;
     } else {
-      (*pRowIndex) += getNumOfRowsInTimeWindow(pDataBlockInfo, startTsCol, *pRowIndex, endWin.ekey, binarySearchForKey,
-                                               NULL, TSDB_ORDER_ASC);
+      while ((groupId == gpIdCol[(*pRowIndex)] && startTsCol[*pRowIndex] < endWin.ekey)) {
+        (*pRowIndex) += 1;
+        if ((*pRowIndex) == pDataBlockInfo->rows) {
+          break;
+        }
+      }
     }
+
     do {
       preWin = endWin;
       getNextTimeWindow(pInterval, &endWin, TSDB_ORDER_ASC);
     } while (endTsCol[(*pRowIndex) - 1] >= endWin.skey);
     endWin = preWin;
-    if (win.ekey == endWin.ekey || (*pRowIndex) == pDataBlockInfo->rows) {
+    if (win.ekey == endWin.ekey || (*pRowIndex) == pDataBlockInfo->rows || groupId != gpIdCol[*pRowIndex]) {
       win.ekey = endWin.ekey;
       return win;
     }
@@ -1235,11 +1245,13 @@ static int32_t generateIntervalScanRange(SStreamScanInfo* pInfo, SSDataBlock* pS
   int64_t          version = pSrcBlock->info.version - 1;
   for (int32_t i = 0; i < rows;) {
     uint64_t srcUid = srcUidData[i];
-    uint64_t groupId = getGroupIdByData(pInfo, srcUid, srcStartTsCol[i], version);
-    uint64_t srcGpId = srcGp[i];
-    TSKEY    calStartTs = srcStartTsCol[i];
+    uint64_t groupId = srcGp[i];
+    if (groupId == 0) {
+      groupId = getGroupIdByData(pInfo, srcUid, srcStartTsCol[i], version);
+    }
+    TSKEY calStartTs = srcStartTsCol[i];
     colDataAppend(pCalStartTsCol, pDestBlock->info.rows, (const char*)(&calStartTs), false);
-    STimeWindow win = getSlidingWindow(srcStartTsCol, srcEndTsCol, &pInfo->interval, &pSrcBlock->info, &i,
+    STimeWindow win = getSlidingWindow(srcStartTsCol, srcEndTsCol, srcGp, &pInfo->interval, &pSrcBlock->info, &i,
                                        pInfo->partitionSup.needCalc);
     TSKEY       calEndTs = srcStartTsCol[i - 1];
     colDataAppend(pCalEndTsCol, pDestBlock->info.rows, (const char*)(&calEndTs), false);
@@ -1248,15 +1260,6 @@ static int32_t generateIntervalScanRange(SStreamScanInfo* pInfo, SSDataBlock* pS
     colDataAppend(pEndTsCol, pDestBlock->info.rows, (const char*)(&win.ekey), false);
     colDataAppend(pGpCol, pDestBlock->info.rows, (const char*)(&groupId), false);
     pDestBlock->info.rows++;
-    if (pInfo->partitionSup.needCalc && srcGpId != 0 && groupId != srcGpId) {
-      colDataAppend(pCalStartTsCol, pDestBlock->info.rows, (const char*)(&calStartTs), false);
-      colDataAppend(pCalEndTsCol, pDestBlock->info.rows, (const char*)(&calEndTs), false);
-      colDataAppend(pDeUidCol, pDestBlock->info.rows, (const char*)(&srcUid), false);
-      colDataAppend(pStartTsCol, pDestBlock->info.rows, (const char*)(&win.skey), false);
-      colDataAppend(pEndTsCol, pDestBlock->info.rows, (const char*)(&win.ekey), false);
-      colDataAppend(pGpCol, pDestBlock->info.rows, (const char*)(&srcGpId), false);
-      pDestBlock->info.rows++;
-    }
   }
   return TSDB_CODE_SUCCESS;
 }
@@ -1331,7 +1334,7 @@ void appendOneRow(SSDataBlock* pBlock, TSKEY* pStartTs, TSKEY* pEndTs, uint64_t*
 static void checkUpdateData(SStreamScanInfo* pInfo, bool invertible, SSDataBlock* pBlock, bool out) {
   if (out) {
     blockDataCleanup(pInfo->pUpdateDataRes);
-    blockDataEnsureCapacity(pInfo->pUpdateDataRes, pBlock->info.rows);
+    blockDataEnsureCapacity(pInfo->pUpdateDataRes, pBlock->info.rows * 2);
   }
   SColumnInfoData* pColDataInfo = taosArrayGet(pBlock->pDataBlock, pInfo->primaryTsIndex);
   ASSERT(pColDataInfo->info.type == TSDB_DATA_TYPE_TIMESTAMP);
@@ -1352,10 +1355,12 @@ static void checkUpdateData(SStreamScanInfo* pInfo, bool invertible, SSDataBlock
                      isDeletedStreamWindow(&win, pBlock->info.groupId, pInfo->pTableScanOp, &pInfo->twAggSup);
     if ((update || closedWin) && out) {
       qDebug("stream update check not pass, update %d, closedWin %d", update, closedWin);
-      uint64_t gpId = closedWin && pInfo->partitionSup.needCalc
-                          ? calGroupIdByData(&pInfo->partitionSup, pInfo->pPartScalarSup, pBlock, rowId)
-                          : 0;
+      uint64_t gpId = 0;
       appendOneRow(pInfo->pUpdateDataRes, tsCol + rowId, tsCol + rowId, &pBlock->info.uid, &gpId);
+      if (closedWin && pInfo->partitionSup.needCalc) {
+        gpId = calGroupIdByData(&pInfo->partitionSup, pInfo->pPartScalarSup, pBlock, rowId);
+        appendOneRow(pInfo->pUpdateDataRes, tsCol + rowId, tsCol + rowId, &pBlock->info.uid, &gpId);
+      }
     }
   }
   if (out && pInfo->pUpdateDataRes->info.rows > 0) {
@@ -1532,6 +1537,30 @@ static int32_t filterDelBlockByUid(SSDataBlock* pDst, const SSDataBlock* pSrc, S
   return 0;
 }
 
+// for partition by tag
+static void setBlockGroupIdByUid(SStreamScanInfo* pInfo, SSDataBlock* pBlock) {
+  SColumnInfoData* pStartTsCol = taosArrayGet(pBlock->pDataBlock, START_TS_COLUMN_INDEX);
+  TSKEY*           startTsCol = (TSKEY*)pStartTsCol->pData;
+  SColumnInfoData* pGpCol = taosArrayGet(pBlock->pDataBlock, GROUPID_COLUMN_INDEX);
+  uint64_t*        gpCol = (uint64_t*)pGpCol->pData;
+  SColumnInfoData* pUidCol = taosArrayGet(pBlock->pDataBlock, UID_COLUMN_INDEX);
+  uint64_t*        uidCol = (uint64_t*)pUidCol->pData;
+  int32_t          rows = pBlock->info.rows;
+  if (!pInfo->partitionSup.needCalc) {
+    for (int32_t i = 0; i < rows; i++) {
+      uint64_t groupId = getGroupIdByUid(pInfo, uidCol[i]);
+      colDataAppend(pGpCol, i, (const char*)&groupId, false);
+    }
+  } else {
+    // SSDataBlock* pPreRes = readPreVersionData(pInfo->pTableScanOp, uidCol[i], startTsCol, ts, maxVersion);
+    // if (!pPreRes || pPreRes->info.rows == 0) {
+    //   return 0;
+    // }
+    // ASSERT(pPreRes->info.rows == 1);
+    // return calGroupIdByData(&pInfo->partitionSup, pInfo->pPartScalarSup, pPreRes, 0);
+  }
+}
+
 static SSDataBlock* doStreamScan(SOperatorInfo* pOperator) {
   // NOTE: this operator does never check if current status is done or not
   SExecTaskInfo*   pTaskInfo = pOperator->pTaskInfo;
@@ -1628,7 +1657,8 @@ FETCH_NEXT_BLOCK:
         } else {
           pDelBlock = pBlock;
         }
-        printDataBlock(pBlock, "stream scan delete recv filtered");
+        setBlockGroupIdByUid(pInfo, pDelBlock);
+        printDataBlock(pDelBlock, "stream scan delete recv filtered");
         if (!isIntervalWindow(pInfo) && !isSessionWindow(pInfo) && !isStateWindow(pInfo)) {
           generateDeleteResultBlock(pInfo, pDelBlock, pInfo->pDeleteDataRes);
           pInfo->pDeleteDataRes->info.type = STREAM_DELETE_RESULT;

source/libs/executor/src/timewindowoperator.c

@@ -271,6 +271,10 @@ static void getNextTimeWindow(SInterval* pInterval, int32_t precision, int32_t o
   tw->ekey -= 1;
 }
 
+void getNextIntervalWindow(SInterval* pInterval, STimeWindow* tw, int32_t order) {
+  getNextTimeWindow(pInterval, pInterval->precision, order, tw);
+}
+
 void doTimeWindowInterpolation(SArray* pPrevValues, SArray* pDataBlock, TSKEY prevTs, int32_t prevRowIndex, TSKEY curTs,
                                int32_t curRowIndex, TSKEY windowKey, int32_t type, SExprSupp* pSup) {
   SqlFunctionCtx* pCtx = pSup->pCtx;
@@ -2095,12 +2099,17 @@ static void genInterpolationResult(STimeSliceOperatorInfo* pSliceInfo, SExprSupp
   bool hasInterp = true;
   for (int32_t j = 0; j < pExprSup->numOfExprs; ++j) {
     SExprInfo* pExprInfo = &pExprSup->pExprInfo[j];
-    int32_t    srcSlot = pExprInfo->base.pParam[0].pCol->slotId;
-    int32_t    dstSlot = pExprInfo->base.resSchema.slotId;
 
-    // SColumnInfoData* pSrc = taosArrayGet(pBlock->pDataBlock, srcSlot);
+    int32_t dstSlot = pExprInfo->base.resSchema.slotId;
     SColumnInfoData* pDst = taosArrayGet(pResBlock->pDataBlock, dstSlot);
 
+    if (IS_TIMESTAMP_TYPE(pExprInfo->base.resSchema.type)) {
+      colDataAppend(pDst, rows, (char*)&pSliceInfo->current, false);
+      continue;
+    }
+
+    int32_t srcSlot = pExprInfo->base.pParam[0].pCol->slotId;
+    // SColumnInfoData* pSrc = taosArrayGet(pBlock->pDataBlock, srcSlot);
     switch (pSliceInfo->fillType) {
       case TSDB_FILL_NULL: {
         colDataAppendNULL(pDst, rows);
@@ -2346,19 +2355,24 @@ static SSDataBlock* doTimeslice(SOperatorInfo* pOperator) {
       if (ts == pSliceInfo->current) {
         for (int32_t j = 0; j < pOperator->exprSupp.numOfExprs; ++j) {
           SExprInfo* pExprInfo = &pOperator->exprSupp.pExprInfo[j];
-          int32_t    dstSlot = pExprInfo->base.resSchema.slotId;
-          int32_t    srcSlot = pExprInfo->base.pParam[0].pCol->slotId;
 
-          SColumnInfoData* pSrc = taosArrayGet(pBlock->pDataBlock, srcSlot);
+          int32_t dstSlot = pExprInfo->base.resSchema.slotId;
           SColumnInfoData* pDst = taosArrayGet(pResBlock->pDataBlock, dstSlot);
 
-          if (colDataIsNull_s(pSrc, i)) {
-            colDataAppendNULL(pDst, pResBlock->info.rows);
-            continue;
-          }
+          if (IS_TIMESTAMP_TYPE(pExprInfo->base.resSchema.type)) {
+            colDataAppend(pDst, pResBlock->info.rows, (char *)&pSliceInfo->current, false);
+          } else {
+            int32_t srcSlot = pExprInfo->base.pParam[0].pCol->slotId;
+            SColumnInfoData* pSrc = taosArrayGet(pBlock->pDataBlock, srcSlot);
+
+            if (colDataIsNull_s(pSrc, i)) {
+              colDataAppendNULL(pDst, pResBlock->info.rows);
+              continue;
+            }
 
-          char* v = colDataGetData(pSrc, i);
-          colDataAppend(pDst, pResBlock->info.rows, v, false);
+            char* v = colDataGetData(pSrc, i);
+            colDataAppend(pDst, pResBlock->info.rows, v, false);
+          }
         }
 
         pResBlock->info.rows += 1;
@@ -2478,14 +2492,24 @@ static SSDataBlock* doTimeslice(SOperatorInfo* pOperator) {
         if (ts == pSliceInfo->current && pSliceInfo->current <= pSliceInfo->win.ekey) {
           for (int32_t j = 0; j < pOperator->exprSupp.numOfExprs; ++j) {
             SExprInfo* pExprInfo = &pOperator->exprSupp.pExprInfo[j];
-            int32_t    dstSlot = pExprInfo->base.resSchema.slotId;
-            int32_t    srcSlot = pExprInfo->base.pParam[0].pCol->slotId;
 
-            SColumnInfoData* pSrc = taosArrayGet(pBlock->pDataBlock, srcSlot);
+            int32_t dstSlot = pExprInfo->base.resSchema.slotId;
             SColumnInfoData* pDst = taosArrayGet(pResBlock->pDataBlock, dstSlot);
 
-            char* v = colDataGetData(pSrc, i);
-            colDataAppend(pDst, pResBlock->info.rows, v, false);
+            if (IS_TIMESTAMP_TYPE(pExprInfo->base.resSchema.type)) {
+              colDataAppend(pDst, pResBlock->info.rows, (char *)&pSliceInfo->current, false);
+            } else {
+              int32_t srcSlot = pExprInfo->base.pParam[0].pCol->slotId;
+              SColumnInfoData* pSrc = taosArrayGet(pBlock->pDataBlock, srcSlot);
+
+              if (colDataIsNull_s(pSrc, i)) {
+                colDataAppendNULL(pDst, pResBlock->info.rows);
+                continue;
+              }
+
+              char* v = colDataGetData(pSrc, i);
+              colDataAppend(pDst, pResBlock->info.rows, v, false);
+            }
           }
 
           pResBlock->info.rows += 1;
@@ -3146,20 +3170,21 @@ static SSDataBlock* doStreamFinalIntervalAgg(SOperatorInfo* pOperator) {
     }
 
     doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
-    if (pInfo->binfo.pRes->info.rows == 0) {
-      pOperator->status = OP_EXEC_DONE;
-      if (!IS_FINAL_OP(pInfo)) {
-        clearFunctionContext(&pOperator->exprSupp);
-        // semi interval operator clear disk buffer
-        clearStreamIntervalOperator(pInfo);
-        qDebug("===stream===clear semi operator");
-      } else {
-        freeAllPages(pInfo->pRecycledPages, pInfo->aggSup.pResultBuf);
-      }
-      return NULL;
+    if (pInfo->binfo.pRes->info.rows != 0) {
+      printDataBlock(pInfo->binfo.pRes, IS_FINAL_OP(pInfo) ? "interval final" : "interval semi");
+      return pInfo->binfo.pRes;
     }
-    printDataBlock(pInfo->binfo.pRes, IS_FINAL_OP(pInfo) ? "interval final" : "interval semi");
-    return pInfo->binfo.pRes;
+
+    doSetOperatorCompleted(pOperator);
+    if (!IS_FINAL_OP(pInfo)) {
+      clearFunctionContext(&pOperator->exprSupp);
+      // semi interval operator clear disk buffer
+      clearStreamIntervalOperator(pInfo);
+      qDebug("===stream===clear semi operator");
+    } else {
+      freeAllPages(pInfo->pRecycledPages, pInfo->aggSup.pResultBuf);
+    }
+    return NULL;
   } else {
     if (!IS_FINAL_OP(pInfo)) {
       doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
@@ -3316,7 +3341,13 @@ static SSDataBlock* doStreamFinalIntervalAgg(SOperatorInfo* pOperator) {
     return pInfo->pPullDataRes;
   }
 
-  // we should send result first.
+  doBuildDeleteResult(pInfo->pDelWins, &pInfo->delIndex, pInfo->pDelRes);
+  if (pInfo->pDelRes->info.rows != 0) {
+    // process the rest of the data
+    printDataBlock(pInfo->pDelRes, IS_FINAL_OP(pInfo) ? "interval final" : "interval semi");
+    return pInfo->pDelRes;
+  }
+
   doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
   if (pInfo->binfo.pRes->info.rows != 0) {
     printDataBlock(pInfo->binfo.pRes, IS_FINAL_OP(pInfo) ? "interval final" : "interval semi");
@@ -3330,13 +3361,6 @@ static SSDataBlock* doStreamFinalIntervalAgg(SOperatorInfo* pOperator) {
     // process the rest of the data
     return pInfo->pUpdateRes;
   }
-
-  doBuildDeleteResult(pInfo->pDelWins, &pInfo->delIndex, pInfo->pDelRes);
-  if (pInfo->pDelRes->info.rows != 0) {
-    // process the rest of the data
-    printDataBlock(pInfo->pDelRes, IS_FINAL_OP(pInfo) ? "interval final" : "interval semi");
-    return pInfo->pDelRes;
-  }
   return NULL;
 }
 
@@ -5744,19 +5768,18 @@ static SSDataBlock* doStreamIntervalAgg(SOperatorInfo* pOperator) {
   if (pOperator->status == OP_RES_TO_RETURN) {
     doBuildDeleteResult(pInfo->pDelWins, &pInfo->delIndex, pInfo->pDelRes);
     if (pInfo->pDelRes->info.rows > 0) {
-      printDataBlock(pInfo->pDelRes, "single interval");
+      printDataBlock(pInfo->pDelRes, "single interval delete");
       return pInfo->pDelRes;
     }
 
     doBuildResult(pOperator, pInfo->binfo.pRes, &pInfo->groupResInfo);
-    // doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
-    if (pInfo->binfo.pRes->info.rows == 0 || !hasRemainResults(&pInfo->groupResInfo)) {
-      pOperator->status = OP_EXEC_DONE;
-      qDebug("===stream===single interval is done");
-      freeAllPages(pInfo->pRecycledPages, pInfo->aggSup.pResultBuf);
+    if (pInfo->binfo.pRes->info.rows > 0) {
+      printDataBlock(pInfo->binfo.pRes, "single interval");
+      return pInfo->binfo.pRes;
     }
-    printDataBlock(pInfo->binfo.pRes, "single interval");
-    return pInfo->binfo.pRes->info.rows == 0 ? NULL : pInfo->binfo.pRes;
+
+    doSetOperatorCompleted(pOperator);
+    return NULL;
   }
 
   SOperatorInfo* downstream = pOperator->pDownstream[0];
@@ -5823,24 +5846,24 @@ static SSDataBlock* doStreamIntervalAgg(SOperatorInfo* pOperator) {
   }
   taosArraySort(pUpdated, resultrowComparAsc);
 
-  // new disc buf
-  // finalizeUpdatedResult(pOperator->exprSupp.numOfExprs, pInfo->aggSup.pResultBuf, pUpdated,
-  // pSup->rowEntryInfoOffset);
   initMultiResInfoFromArrayList(&pInfo->groupResInfo, pUpdated);
   blockDataEnsureCapacity(pInfo->binfo.pRes, pOperator->resultInfo.capacity);
   removeDeleteResults(pUpdatedMap, pInfo->pDelWins);
   taosHashCleanup(pUpdatedMap);
+
   doBuildDeleteResult(pInfo->pDelWins, &pInfo->delIndex, pInfo->pDelRes);
   if (pInfo->pDelRes->info.rows > 0) {
-    printDataBlock(pInfo->pDelRes, "single interval");
+    printDataBlock(pInfo->pDelRes, "single interval delete");
     return pInfo->pDelRes;
   }
 
-  // doBuildResultDatablock(pOperator, &pInfo->binfo, &pInfo->groupResInfo, pInfo->aggSup.pResultBuf);
-  // new disc buf
   doBuildResult(pOperator, pInfo->binfo.pRes, &pInfo->groupResInfo);
-  printDataBlock(pInfo->binfo.pRes, "single interval");
-  return pInfo->binfo.pRes->info.rows == 0 ? NULL : pInfo->binfo.pRes;
+  if (pInfo->binfo.pRes->info.rows > 0) {
+    printDataBlock(pInfo->binfo.pRes, "single interval");
+    return pInfo->binfo.pRes;
+  }
+
+  return NULL;
 }
 
 void destroyStreamIntervalOperatorInfo(void* param) {

source/libs/function/inc/functionMgtInt.h

@@ -50,6 +50,7 @@ extern "C" {
 #define FUNC_MGT_KEEP_ORDER_FUNC        FUNC_MGT_FUNC_CLASSIFICATION_MASK(21)
 #define FUNC_MGT_CUMULATIVE_FUNC        FUNC_MGT_FUNC_CLASSIFICATION_MASK(22)
 #define FUNC_MGT_FORBID_STABLE_FUNC     FUNC_MGT_FUNC_CLASSIFICATION_MASK(23)
+#define FUNC_MGT_INTERP_PC_FUNC         FUNC_MGT_FUNC_CLASSIFICATION_MASK(24)
 
 #define FUNC_MGT_TEST_MASK(val, mask) (((val) & (mask)) != 0)
 

source/libs/function/src/builtins.c

@@ -3146,6 +3146,16 @@ const SBuiltinFuncDefinition funcMgtBuiltins[] = {
     .classification = FUNC_MGT_SYSTEM_INFO_FUNC | FUNC_MGT_SCALAR_FUNC,
     .translateFunc = translateUserFunc,
   },
+  {
+    .name = "_irowts",
+    .type = FUNCTION_TYPE_IROWTS,
+    .classification = FUNC_MGT_PSEUDO_COLUMN_FUNC | FUNC_MGT_INTERP_PC_FUNC,
+    .translateFunc = translateTimePseudoColumn,
+    .getEnvFunc   = getTimePseudoFuncEnv,
+    .initFunc     = NULL,
+    .sprocessFunc = NULL,
+    .finalizeFunc = NULL
+  },
 };
 // clang-format on
 

source/libs/function/src/functionMgt.c

@@ -224,6 +224,8 @@ bool fmIsInterpFunc(int32_t funcId) {
   return FUNCTION_TYPE_INTERP == funcMgtBuiltins[funcId].type;
 }
 
+bool fmIsInterpPseudoColumnFunc(int32_t funcId) { return isSpecificClassifyFunc(funcId, FUNC_MGT_INTERP_PC_FUNC); }
+
 bool fmIsLastRowFunc(int32_t funcId) {
   if (funcId < 0 || funcId >= funcMgtBuiltinsNum) {
     return false;

source/libs/index/src/indexCache.c

@@ -302,6 +302,7 @@ static int32_t cacheSearchCompareFunc_JSON(void* cache, SIndexTerm* term, SIdxTR
         char* p = taosMemoryCalloc(1, strlen(c->colVal) + 1);
         memcpy(p, c->colVal, strlen(c->colVal));
         cond = cmpFn(p + skip, term->colVal, dType);
+        taosMemoryFree(p);
       }
     }
     if (cond == MATCH) {

source/libs/index/src/indexFstFile.c

@@ -69,6 +69,8 @@ static int idxFileCtxDoReadFrom(IFileCtx* ctx, uint8_t* buf, int len, int32_t of
   int32_t blkOffset = offset % kBlockSize;
   int32_t blkLeft = kBlockSize - blkOffset;
 
+  if (offset >= ctx->file.size) return 0;
+
   do {
     char key[128] = {0};
     idxGenLRUKey(key, ctx->file.buf, blkId);
@@ -80,24 +82,34 @@ static int idxFileCtxDoReadFrom(IFileCtx* ctx, uint8_t* buf, int len, int32_t of
       memcpy(buf + total, blk->buf + blkOffset, nread);
       taosLRUCacheRelease(ctx->lru, h, false);
     } else {
-      int32_t cacheMemSize = sizeof(SDataBlock) + kBlockSize;
+      int32_t left = ctx->file.size - offset;
+      if (left < kBlockSize) {
+        nread = TMIN(left, len);
+        int32_t bytes = taosPReadFile(ctx->file.pFile, buf + total, nread, offset);
+        assert(bytes == nread);
 
-      SDataBlock* blk = taosMemoryCalloc(1, cacheMemSize);
-      blk->blockId = blkId;
-      blk->nread = taosPReadFile(ctx->file.pFile, blk->buf, kBlockSize, blkId * kBlockSize);
-      assert(blk->nread <= kBlockSize);
+        total += bytes;
+        return total;
+      } else {
+        int32_t cacheMemSize = sizeof(SDataBlock) + kBlockSize;
 
-      if (blk->nread < kBlockSize && blk->nread < len) {
-        break;
-      }
+        SDataBlock* blk = taosMemoryCalloc(1, cacheMemSize);
+        blk->blockId = blkId;
+        blk->nread = taosPReadFile(ctx->file.pFile, blk->buf, kBlockSize, blkId * kBlockSize);
+        assert(blk->nread <= kBlockSize);
 
-      nread = TMIN(blkLeft, len);
-      memcpy(buf + total, blk->buf + blkOffset, nread);
+        if (blk->nread < kBlockSize && blk->nread < len) {
+          break;
+        }
+
+        nread = TMIN(blkLeft, len);
+        memcpy(buf + total, blk->buf + blkOffset, nread);
 
-      LRUStatus s = taosLRUCacheInsert(ctx->lru, key, strlen(key), blk, cacheMemSize, deleteDataBlockFromLRU, NULL,
-                                       TAOS_LRU_PRIORITY_LOW);
-      if (s != TAOS_LRU_STATUS_OK) {
-        return -1;
+        LRUStatus s = taosLRUCacheInsert(ctx->lru, key, strlen(key), blk, cacheMemSize, deleteDataBlockFromLRU, NULL,
+                                         TAOS_LRU_PRIORITY_LOW);
+        if (s != TAOS_LRU_STATUS_OK) {
+          return -1;
+        }
       }
     }
     total += nread;
@@ -146,9 +158,7 @@ IFileCtx* idxFileCtxCreate(WriterType type, const char* path, bool readOnly, int
     } else {
       ctx->file.pFile = taosOpenFile(path, TD_FILE_READ);
 
-      int64_t size = 0;
       taosFStatFile(ctx->file.pFile, &ctx->file.size, NULL);
-      ctx->file.size = (int)size;
 #ifdef USE_MMAP
       ctx->file.ptr = (char*)tfMmapReadOnly(ctx->file.pFile, ctx->file.size);
 #endif

source/libs/index/test/jsonUT.cc

@@ -172,9 +172,9 @@ TEST_F(JsonEnv, testWriteMillonData) {
   {
     std::string colName("voltagefdadfa");
     std::string colVal("abxxxxxxxxxxxx");
-    for (int i = 0; i < 10; i++) {
+    for (int i = 0; i < 10000; i++) {
       colVal[i % colVal.size()] = '0' + i % 128;
-      for (size_t i = 0; i < 100; i++) {
+      for (size_t i = 0; i < 10; i++) {
         SIndexTerm* term = indexTermCreateT(1, ADD_VALUE, TSDB_DATA_TYPE_BINARY, colName.c_str(), colName.size(),
                                             colVal.c_str(), colVal.size());
 

source/libs/nodes/src/nodesCodeFuncs.c

@@ -254,6 +254,7 @@ const char* nodesNodeName(ENodeType type) {
     case QUERY_NODE_PHYSICAL_PLAN_STREAM_SEMI_INTERVAL:
       return "PhysiStreamSemiInterval";
     case QUERY_NODE_PHYSICAL_PLAN_FILL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL:
       return "PhysiFill";
     case QUERY_NODE_PHYSICAL_PLAN_MERGE_SESSION:
       return "PhysiSessionWindow";
@@ -4635,6 +4636,7 @@ static int32_t specificNodeToJson(const void* pObj, SJson* pJson) {
     case QUERY_NODE_PHYSICAL_PLAN_STREAM_SEMI_INTERVAL:
       return physiIntervalNodeToJson(pObj, pJson);
     case QUERY_NODE_PHYSICAL_PLAN_FILL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL:
       return physiFillNodeToJson(pObj, pJson);
     case QUERY_NODE_PHYSICAL_PLAN_MERGE_SESSION:
     case QUERY_NODE_PHYSICAL_PLAN_STREAM_SESSION:
@@ -4788,6 +4790,7 @@ static int32_t jsonToSpecificNode(const SJson* pJson, void* pObj) {
     case QUERY_NODE_PHYSICAL_PLAN_STREAM_SEMI_INTERVAL:
       return jsonToPhysiIntervalNode(pJson, pObj);
     case QUERY_NODE_PHYSICAL_PLAN_FILL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL:
       return jsonToPhysiFillNode(pJson, pObj);
     case QUERY_NODE_PHYSICAL_PLAN_MERGE_SESSION:
     case QUERY_NODE_PHYSICAL_PLAN_STREAM_SESSION:

source/libs/nodes/src/nodesMsgFuncs.c

@@ -3633,6 +3633,7 @@ static int32_t specificNodeToMsg(const void* pObj, STlvEncoder* pEncoder) {
       code = physiIntervalNodeToMsg(pObj, pEncoder);
       break;
     case QUERY_NODE_PHYSICAL_PLAN_FILL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL:
       code = physiFillNodeToMsg(pObj, pEncoder);
       break;
     case QUERY_NODE_PHYSICAL_PLAN_MERGE_SESSION:
@@ -3770,6 +3771,7 @@ static int32_t msgToSpecificNode(STlvDecoder* pDecoder, void* pObj) {
       code = msgToPhysiIntervalNode(pDecoder, pObj);
       break;
     case QUERY_NODE_PHYSICAL_PLAN_FILL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL:
       code = msgToPhysiFillNode(pDecoder, pObj);
       break;
     case QUERY_NODE_PHYSICAL_PLAN_MERGE_SESSION:

source/libs/nodes/src/nodesUtilFuncs.c

@@ -511,6 +511,7 @@ SNode* nodesMakeNode(ENodeType type) {
     case QUERY_NODE_PHYSICAL_PLAN_STREAM_SEMI_INTERVAL:
       return makeNode(type, sizeof(SStreamSemiIntervalPhysiNode));
     case QUERY_NODE_PHYSICAL_PLAN_FILL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL:
       return makeNode(type, sizeof(SFillPhysiNode));
     case QUERY_NODE_PHYSICAL_PLAN_MERGE_SESSION:
       return makeNode(type, sizeof(SSessionWinodwPhysiNode));
@@ -1156,7 +1157,8 @@ void nodesDestroyNode(SNode* pNode) {
     case QUERY_NODE_PHYSICAL_PLAN_STREAM_SEMI_INTERVAL:
       destroyWinodwPhysiNode((SWinodwPhysiNode*)pNode);
       break;
-    case QUERY_NODE_PHYSICAL_PLAN_FILL: {
+    case QUERY_NODE_PHYSICAL_PLAN_FILL:
+    case QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL: {
       SFillPhysiNode* pPhyNode = (SFillPhysiNode*)pNode;
       destroyPhysiNode((SPhysiNode*)pPhyNode);
       nodesDestroyList(pPhyNode->pFillExprs);

source/libs/parser/inc/sql.y

@@ -707,6 +707,7 @@ pseudo_column(A) ::= QDURATION(B).
 pseudo_column(A) ::= WSTART(B).                                                   { A = createRawExprNode(pCxt, &B, createFunctionNode(pCxt, &B, NULL)); }
 pseudo_column(A) ::= WEND(B).                                                     { A = createRawExprNode(pCxt, &B, createFunctionNode(pCxt, &B, NULL)); }
 pseudo_column(A) ::= WDURATION(B).                                                { A = createRawExprNode(pCxt, &B, createFunctionNode(pCxt, &B, NULL)); }
+pseudo_column(A) ::= IROWTS(B).                                                   { A = createRawExprNode(pCxt, &B, createFunctionNode(pCxt, &B, NULL)); }
 
 function_expression(A) ::= function_name(B) NK_LP expression_list(C) NK_RP(D).    { A = createRawExprNodeExt(pCxt, &B, &D, createFunctionNode(pCxt, &B, C)); }
 function_expression(A) ::= star_func(B) NK_LP star_func_para_list(C) NK_RP(D).    { A = createRawExprNodeExt(pCxt, &B, &D, createFunctionNode(pCxt, &B, C)); }

source/libs/parser/src/parAstParser.c

@@ -407,6 +407,10 @@ static int32_t collectMetaKeyFromShowTags(SCollectMetaKeyCxt* pCxt, SShowStmt* p
   if (TSDB_CODE_SUCCESS == code) {
     code = reserveDbVgInfoInCache(pCxt->pParseCxt->acctId, ((SValueNode*)pStmt->pDbName)->literal, pCxt->pMetaCache);
   }
+  if (TSDB_CODE_SUCCESS == code && NULL != pStmt->pTbName) {
+    code = reserveTableVgroupInCache(pCxt->pParseCxt->acctId, ((SValueNode*)pStmt->pDbName)->literal,
+                                     ((SValueNode*)pStmt->pTbName)->literal, pCxt->pMetaCache);
+  }
   return code;
 }
 

source/libs/parser/src/parTokenizer.c

@@ -252,6 +252,7 @@ static SKeyword keywordTable[] = {
     {"WITH",                 TK_WITH},
     {"WRITE",                TK_WRITE},
     {"_C0",                  TK_ROWTS},
+    {"_IROWTS",              TK_IROWTS},
     {"_QDURATION",           TK_QDURATION},
     {"_QEND",                TK_QEND},
     {"_QSTART",              TK_QSTART},

source/libs/parser/src/parTranslater.c

@@ -1556,6 +1556,9 @@ static int32_t translateMultiResFunc(STranslateContext* pCxt, SFunctionNode* pFu
                                      "%s(*) is only supported in SELECTed list", pFunc->functionName);
     }
   }
+  if (tsKeepColumnName) {
+    strcpy(pFunc->node.userAlias, ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 0))->userAlias);
+  }
   return TSDB_CODE_SUCCESS;
 }
 
@@ -2138,10 +2141,114 @@ static bool sysTableFromVnode(const char* pTable) {
 
 static bool sysTableFromDnode(const char* pTable) { return 0 == strcmp(pTable, TSDB_INS_TABLE_DNODE_VARIABLES); }
 
+static int32_t getTagsTableVgroupListImpl(STranslateContext* pCxt, SName* pTargetName, SName* pName,
+                                          SArray** pVgroupList) {
+  if (0 == pTargetName->type) {
+    return getDBVgInfoImpl(pCxt, pName, pVgroupList);
+  }
+
+  if (TSDB_DB_NAME_T == pTargetName->type) {
+    return getDBVgInfoImpl(pCxt, pTargetName, pVgroupList);
+  }
+
+  SVgroupInfo vgInfo = {0};
+  int32_t     code = getTableHashVgroupImpl(pCxt, pTargetName, &vgInfo);
+  if (TSDB_CODE_SUCCESS == code) {
+    *pVgroupList = taosArrayInit(1, sizeof(SVgroupInfo));
+    if (NULL == *pVgroupList) {
+      code = TSDB_CODE_OUT_OF_MEMORY;
+    }
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    taosArrayPush(*pVgroupList, &vgInfo);
+  }
+  return code;
+}
+
+static int32_t getTagsTableTargetNameFromOp(STranslateContext* pCxt, SOperatorNode* pOper, SName* pName) {
+  if (OP_TYPE_EQUAL != pOper->opType) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  SColumnNode* pCol = NULL;
+  SValueNode*  pVal = NULL;
+  if (QUERY_NODE_COLUMN == nodeType(pOper->pLeft)) {
+    pCol = (SColumnNode*)pOper->pLeft;
+  } else if (QUERY_NODE_VALUE == nodeType(pOper->pLeft)) {
+    pVal = (SValueNode*)pOper->pLeft;
+  }
+  if (QUERY_NODE_COLUMN == nodeType(pOper->pRight)) {
+    pCol = (SColumnNode*)pOper->pRight;
+  } else if (QUERY_NODE_VALUE == nodeType(pOper->pRight)) {
+    pVal = (SValueNode*)pOper->pRight;
+  }
+  if (NULL == pCol || NULL == pVal) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  if (0 == strcmp(pCol->colName, "db_name")) {
+    return tNameSetDbName(pName, pCxt->pParseCxt->acctId, pVal->literal, strlen(pVal->literal));
+  } else if (0 == strcmp(pCol->colName, "table_name")) {
+    return tNameAddTbName(pName, pVal->literal, strlen(pVal->literal));
+  }
+
+  return TSDB_CODE_SUCCESS;
+}
+
+static void getTagsTableTargetObjName(STranslateContext* pCxt, SNode* pNode, SName* pName) {
+  if (QUERY_NODE_OPERATOR == nodeType(pNode)) {
+    getTagsTableTargetNameFromOp(pCxt, (SOperatorNode*)pNode, pName);
+  }
+}
+
+static int32_t getTagsTableTargetNameFromCond(STranslateContext* pCxt, SLogicConditionNode* pCond, SName* pName) {
+  if (LOGIC_COND_TYPE_AND != pCond->condType) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  SNode* pNode = NULL;
+  FOREACH(pNode, pCond->pParameterList) { getTagsTableTargetObjName(pCxt, pNode, pName); }
+  if ('\0' == pName->dbname[0]) {
+    pName->type = 0;
+  }
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t getTagsTableTargetName(STranslateContext* pCxt, SNode* pWhere, SName* pName) {
+  if (NULL == pWhere) {
+    return TSDB_CODE_SUCCESS;
+  }
+
+  if (QUERY_NODE_OPERATOR == nodeType(pWhere)) {
+    return getTagsTableTargetNameFromOp(pCxt, (SOperatorNode*)pWhere, pName);
+  }
+
+  if (QUERY_NODE_LOGIC_CONDITION == nodeType(pWhere)) {
+    return getTagsTableTargetNameFromCond(pCxt, (SLogicConditionNode*)pWhere, pName);
+  }
+
+  return TSDB_CODE_SUCCESS;
+}
+
+static int32_t getTagsTableVgroupList(STranslateContext* pCxt, SName* pName, SArray** pVgroupList) {
+  if (!isSelectStmt(pCxt->pCurrStmt)) {
+    return TSDB_CODE_SUCCESS;
+  }
+  SSelectStmt* pSelect = (SSelectStmt*)pCxt->pCurrStmt;
+  SName        targetName = {0};
+  int32_t      code = getTagsTableTargetName(pCxt, pSelect->pWhere, &targetName);
+  if (TSDB_CODE_SUCCESS == code) {
+    code = getTagsTableVgroupListImpl(pCxt, &targetName, pName, pVgroupList);
+  }
+  return code;
+}
+
 static int32_t setVnodeSysTableVgroupList(STranslateContext* pCxt, SName* pName, SRealTableNode* pRealTable) {
   int32_t code = TSDB_CODE_SUCCESS;
   SArray* vgroupList = NULL;
-  if ('\0' != pRealTable->qualDbName[0]) {
+  if (0 == strcmp(pRealTable->table.tableName, TSDB_INS_TABLE_TAGS)) {
+    code = getTagsTableVgroupList(pCxt, pName, &vgroupList);
+  } else if ('\0' != pRealTable->qualDbName[0]) {
     if (0 != strcmp(pRealTable->qualDbName, TSDB_INFORMATION_SCHEMA_DB)) {
       code = getDBVgInfo(pCxt, pRealTable->qualDbName, &vgroupList);
     }
@@ -2149,14 +2256,12 @@ static int32_t setVnodeSysTableVgroupList(STranslateContext* pCxt, SName* pName,
     code = getDBVgInfoImpl(pCxt, pName, &vgroupList);
   }
 
-  if (TSDB_CODE_SUCCESS == code && 0 == strcmp(pRealTable->table.dbName, TSDB_INFORMATION_SCHEMA_DB) &&
-      0 == strcmp(pRealTable->table.tableName, TSDB_INS_TABLE_TAGS) && isSelectStmt(pCxt->pCurrStmt) &&
-      0 == taosArrayGetSize(vgroupList)) {
+  if (TSDB_CODE_SUCCESS == code && 0 == strcmp(pRealTable->table.tableName, TSDB_INS_TABLE_TAGS) &&
+      isSelectStmt(pCxt->pCurrStmt) && 0 == taosArrayGetSize(vgroupList)) {
     ((SSelectStmt*)pCxt->pCurrStmt)->isEmptyResult = true;
   }
 
-  if (TSDB_CODE_SUCCESS == code && 0 == strcmp(pRealTable->table.dbName, TSDB_INFORMATION_SCHEMA_DB) &&
-      0 == strcmp(pRealTable->table.tableName, TSDB_INS_TABLE_TABLES)) {
+  if (TSDB_CODE_SUCCESS == code && 0 == strcmp(pRealTable->table.tableName, TSDB_INS_TABLE_TABLES)) {
     code = addMnodeToVgroupList(&pCxt->pParseCxt->mgmtEpSet, &vgroupList);
   }
 
@@ -2388,8 +2493,12 @@ static SNode* createMultiResFunc(SFunctionNode* pSrcFunc, SExprNode* pExpr) {
     SColumnNode* pCol = (SColumnNode*)pExpr;
     len = snprintf(buf, sizeof(buf), "%s(%s.%s)", pSrcFunc->functionName, pCol->tableAlias, pCol->colName);
     strncpy(pFunc->node.aliasName, buf, TMIN(len, sizeof(pFunc->node.aliasName) - 1));
-    len = snprintf(buf, sizeof(buf), "%s(%s)", pSrcFunc->functionName, pCol->colName);
-    strncpy(pFunc->node.userAlias, buf, TMIN(len, sizeof(pFunc->node.userAlias) - 1));
+    if (tsKeepColumnName) {
+      strcpy(pFunc->node.userAlias, pCol->colName);
+    } else {
+      len = snprintf(buf, sizeof(buf), "%s(%s)", pSrcFunc->functionName, pCol->colName);
+      strncpy(pFunc->node.userAlias, buf, TMIN(len, sizeof(pFunc->node.userAlias) - 1));
+    }
   } else {
     len = snprintf(buf, sizeof(buf), "%s(%s)", pSrcFunc->functionName, pExpr->aliasName);
     strncpy(pFunc->node.aliasName, buf, TMIN(len, sizeof(pFunc->node.aliasName) - 1));

source/libs/parser/test/mockCatalog.cpp

@@ -85,6 +85,10 @@ void generateInformationSchema(MockCatalogService* mcs) {
       .addColumn("dnode_id", TSDB_DATA_TYPE_INT)
       .addColumn("dnode_ep", TSDB_DATA_TYPE_BINARY, TSDB_EP_LEN)
       .done();
+  mcs->createTableBuilder(TSDB_INFORMATION_SCHEMA_DB, TSDB_INS_TABLE_TAGS, TSDB_SYSTEM_TABLE, 2)
+      .addColumn("table_name", TSDB_DATA_TYPE_BINARY, TSDB_TABLE_NAME_LEN)
+      .addColumn("db_name", TSDB_DATA_TYPE_BINARY, TSDB_DB_NAME_LEN)
+      .done();
 }
 
 void generatePerformanceSchema(MockCatalogService* mcs) {

source/libs/parser/test/parShowToUse.cpp

@@ -196,6 +196,12 @@ TEST_F(ParserShowToUseTest, showTableDistributed) {
   run("SHOW TABLE DISTRIBUTED st1");
 }
 
+TEST_F(ParserShowToUseTest, showTags) {
+  useDb("root", "test");
+
+  run("SHOW TAGS FROM st1s1");
+}
+
 // todo SHOW topics
 
 TEST_F(ParserShowToUseTest, showUsers) {
@@ -213,9 +219,9 @@ TEST_F(ParserShowToUseTest, showVariables) {
 TEST_F(ParserShowToUseTest, showVgroups) {
   useDb("root", "test");
 
-  run("SHOW vgroups");
+  run("SHOW VGROUPS");
 
-  run("SHOW test.vgroups");
+  run("SHOW test.VGROUPS");
 }
 
 TEST_F(ParserShowToUseTest, showVnodes) {

source/libs/parser/test/parTestMain.cpp

@@ -86,6 +86,7 @@ static void parseArg(int argc, char* argv[]) {
     {"dump", no_argument, NULL, 'd'},
     {"async", required_argument, NULL, 'a'},
     {"skipSql", required_argument, NULL, 's'},
+    {"log", required_argument, NULL, 'l'},
     {0, 0, 0, 0}
   };
   // clang-format on
@@ -100,6 +101,9 @@ static void parseArg(int argc, char* argv[]) {
       case 's':
         setSkipSqlNum(optarg);
         break;
+      case 'l':
+        setLogLevel(optarg);
+        break;
       default:
         break;
     }

source/libs/planner/src/planLogicCreater.c

@@ -611,6 +611,8 @@ static int32_t createIndefRowsFuncLogicNode(SLogicPlanContext* pCxt, SSelectStmt
   return code;
 }
 
+static bool isInterpFunc(int32_t funcId) { return fmIsInterpFunc(funcId) || fmIsInterpPseudoColumnFunc(funcId); }
+
 static int32_t createInterpFuncLogicNode(SLogicPlanContext* pCxt, SSelectStmt* pSelect, SLogicNode** pLogicNode) {
   if (!pSelect->hasInterpFunc) {
     return TSDB_CODE_SUCCESS;
@@ -625,7 +627,7 @@ static int32_t createInterpFuncLogicNode(SLogicPlanContext* pCxt, SSelectStmt* p
   pInterpFunc->node.requireDataOrder = getRequireDataOrder(true, pSelect);
   pInterpFunc->node.resultDataOrder = pInterpFunc->node.requireDataOrder;
 
-  int32_t code = nodesCollectFuncs(pSelect, SQL_CLAUSE_SELECT, fmIsInterpFunc, &pInterpFunc->pFuncs);
+  int32_t code = nodesCollectFuncs(pSelect, SQL_CLAUSE_SELECT, isInterpFunc, &pInterpFunc->pFuncs);
   if (TSDB_CODE_SUCCESS == code) {
     code = rewriteExprsForSelect(pInterpFunc->pFuncs, pSelect, SQL_CLAUSE_SELECT);
   }

source/libs/planner/src/planPhysiCreater.c

@@ -1409,7 +1409,9 @@ static int32_t createPartitionPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChi
 
 static int32_t createFillPhysiNode(SPhysiPlanContext* pCxt, SNodeList* pChildren, SFillLogicNode* pFillNode,
                                    SPhysiNode** pPhyNode) {
-  SFillPhysiNode* pFill = (SFillPhysiNode*)makePhysiNode(pCxt, (SLogicNode*)pFillNode, QUERY_NODE_PHYSICAL_PLAN_FILL);
+  SFillPhysiNode* pFill = (SFillPhysiNode*)makePhysiNode(
+      pCxt, (SLogicNode*)pFillNode,
+      pCxt->pPlanCxt->streamQuery ? QUERY_NODE_PHYSICAL_PLAN_STREAM_FILL : QUERY_NODE_PHYSICAL_PLAN_FILL);
   if (NULL == pFill) {
     return TSDB_CODE_OUT_OF_MEMORY;
   }

source/libs/planner/test/planBasicTest.cpp

@@ -101,6 +101,8 @@ TEST_F(PlanBasicTest, interpFunc) {
   useDb("root", "test");
 
   run("SELECT INTERP(c1) FROM t1 RANGE('2017-7-14 18:00:00', '2017-7-14 19:00:00') EVERY(5s) FILL(LINEAR)");
+
+  run("SELECT _IROWTS, INTERP(c1) FROM t1 RANGE('2017-7-14 18:00:00', '2017-7-14 19:00:00') EVERY(5s) FILL(LINEAR)");
 }
 
 TEST_F(PlanBasicTest, lastRowFunc) {

source/libs/planner/test/planOtherTest.cpp

@@ -84,6 +84,8 @@ TEST_F(PlanOtherTest, show) {
   run("SHOW TABLE DISTRIBUTED st1");
 
   run("SHOW DNODE 1 VARIABLES");
+
+  run("SHOW TAGS FROM st1s1");
 }
 
 TEST_F(PlanOtherTest, delete) {

source/libs/qworker/src/qworker.c

@@ -107,9 +107,12 @@ int32_t qwExecTask(QW_FPARAMS_DEF, SQWTaskCtx *ctx, bool *queryStop) {
     QW_TASK_DLOG("start to execTask, loopIdx:%d", i++);
 
     // if *taskHandle is NULL, it's killed right now
+    bool hasMore = false;
+
     if (taskHandle) {
       qwDbgSimulateSleep();
-      code = qExecTaskOpt(taskHandle, pResList, &useconds, &localFetch);
+
+      code = qExecTaskOpt(taskHandle, pResList, &useconds, &hasMore, &localFetch);
       if (code) {
         if (code != TSDB_CODE_OPS_NOT_SUPPORT) {
           QW_TASK_ELOG("qExecTask failed, code:%x - %s", code, tstrerror(code));
@@ -122,20 +125,8 @@ int32_t qwExecTask(QW_FPARAMS_DEF, SQWTaskCtx *ctx, bool *queryStop) {
 
     ++execNum;
 
-    if (taosArrayGetSize(pResList) == 0) {
-      QW_TASK_DLOG("qExecTask end with empty res, useconds:%" PRIu64, useconds);
-      dsEndPut(sinkHandle, useconds);
-
-      QW_ERR_JRET(qwHandleTaskComplete(QW_FPARAMS(), ctx));
-
-      if (queryStop) {
-        *queryStop = true;
-      }
-
-      break;
-    }
-
-    for (int32_t j = 0; j < taosArrayGetSize(pResList); ++j) {
+    size_t numOfResBlock = taosArrayGetSize(pResList);
+    for (int32_t j = 0; j < numOfResBlock; ++j) {
       SSDataBlock *pRes = taosArrayGetP(pResList, j);
       ASSERT(pRes->info.rows > 0);
 
@@ -149,6 +140,23 @@ int32_t qwExecTask(QW_FPARAMS_DEF, SQWTaskCtx *ctx, bool *queryStop) {
       QW_TASK_DLOG("data put into sink, rows:%d, continueExecTask:%d", pRes->info.rows, qcontinue);
     }
 
+    if (numOfResBlock == 0 || (hasMore == false)) {
+      if (numOfResBlock == 0) {
+        QW_TASK_DLOG("qExecTask end with empty res, useconds:%" PRIu64, useconds);
+      } else {
+        QW_TASK_DLOG("qExecTask done", "");
+      }
+
+      dsEndPut(sinkHandle, useconds);
+      QW_ERR_JRET(qwHandleTaskComplete(QW_FPARAMS(), ctx));
+
+      if (queryStop) {
+        *queryStop = true;
+      }
+
+      break;
+    }
+
     if (!qcontinue) {
       if (queryStop) {
         *queryStop = true;

source/libs/scheduler/src/schJob.c

@@ -47,8 +47,7 @@ void schUpdateJobErrCode(SSchJob *pJob, int32_t errCode) {
   return;
 
 _return:
-
-  SCH_JOB_DLOG("job errCode updated to %x - %s", errCode, tstrerror(errCode));
+  SCH_JOB_DLOG("job errCode updated to %s", tstrerror(errCode));
 }
 
 bool schJobDone(SSchJob *pJob) {
@@ -491,7 +490,7 @@ int32_t schProcessOnJobFailure(SSchJob *pJob, int32_t errCode) {
   
   int32_t code = atomic_load_32(&pJob->errCode);
   if (code) {
-    SCH_JOB_DLOG("job failed with error: %s", tstrerror(code));
+    SCH_JOB_DLOG("job failed with error %s", tstrerror(code));
   }
 
   schPostJobRes(pJob, 0);