Merge branch '3.0' into feature/TD-14481-3.0

Jenkinsfile2

@@ -269,7 +269,7 @@ pipeline {
                     }
                 }
                 stage('linux test') {
-                    agent{label " slave3_0 || slave15 || slave16 || slave17 "}
+                    agent{label " worker03 || slave215 || slave217 || slave219 "}
                     options { skipDefaultCheckout() } 
                     when {
                         changeRequest()
@@ -287,9 +287,9 @@ pipeline {
                                 '''
                                 sh '''
                                     cd ${WKC}/tests/parallel_test
-                                    export DEFAULT_RETRY_TIME=1
+                                    export DEFAULT_RETRY_TIME=2
                                     date
-                                    timeout 2100 time ./run.sh -e -m /home/m.json -t /tmp/cases.task -b ${BRANCH_NAME} -l ${WKDIR}/log -o 480
+                                    timeout 2100 time ./run.sh -e -m /home/m.json -t /tmp/cases.task -b ${BRANCH_NAME}_${BUILD_ID} -l ${WKDIR}/log -o 480
                                 '''
                             }
                         }

docs-cn/12-taos-sql/index.md

@@ -7,8 +7,6 @@ description: "TAOS SQL 支持的语法规则、主要查询功能、支持的 SQ
 
 TAOS SQL 是用户对 TDengine 进行数据写入和查询的主要工具。TAOS SQL 为了便于用户快速上手，在一定程度上提供与标准 SQL 类似的风格和模式。严格意义上，TAOS SQL 并不是也不试图提供标准的 SQL 语法。此外，由于 TDengine 针对的时序性结构化数据不提供删除功能，因此在 TAO SQL 中不提供数据删除的相关功能。
 
-TAOS SQL 不支持关键字的缩写，例如 DESCRIBE 不能缩写为 DESC。
-
 本章节 SQL 语法遵循如下约定：
 
 - <\> 里的内容是用户需要输入的，但不要输入 <\> 本身
@@ -37,4 +35,4 @@ import DocCardList from '@theme/DocCardList';
 import {useCurrentSidebarCategory} from '@docusaurus/theme-common';
 
 <DocCardList items={useCurrentSidebarCategory().items}/>
-```
+```
\ No newline at end of file

docs-en/12-taos-sql/04-stable.md

@@ -9,7 +9,7 @@ Keyword `STable`, abbreviated for super table, is supported since version 2.0.15
 
 :::
 
-## Crate STable
+## Create STable
 
 ```
 CREATE STable [IF NOT EXISTS] stb_name (timestamp_field_name TIMESTAMP, field1_name data_type1 [, field2_name data_type2 ...]) TAGS (tag1_name tag_type1, tag2_name tag_type2 [, tag3_name tag_type3]);
@@ -19,7 +19,7 @@ The SQL statement of creating a STable is similar to that of creating a table, b
 
 :::info
 
-1. The tag types specified in TAGS should NOT be timestamp. Since 2.1.3.0 timestamp type can be used in TAGS column, but its value must be fixed and arithmetic operation can't be applied on it.
+1. A tag can be of type timestamp, since version 2.1.3.0, but its value must be fixed and arithmetic operations cannot be performed on it. Prior to version 2.1.3.0, tag types specified in TAGS could not be of type timestamp.
 2. The tag names specified in TAGS should NOT be the same as other columns.
 3. The tag names specified in TAGS should NOT be the same as any reserved keywords.(Please refer to [keywords](/taos-sql/keywords/)
 4. The maximum number of tags specified in TAGS is 128, there must be at least one tag, and the total length of all tag columns should NOT exceed 16KB.
@@ -76,7 +76,7 @@ ALTER STable stb_name DROP COLUMN field_name;
 ALTER STable stb_name MODIFY COLUMN field_name data_type(length);
 ```
 
-This command can be used to change (or increase, more specifically) the length of a column of variable length types, like BINARY or NCHAR.
+This command can be used to change (or more specifically, increase) the length of a column of variable length types, like BINARY or NCHAR.
 
 ## Change Tags of A STable
 
@@ -94,7 +94,7 @@ This command is used to add a new tag for a STable and specify the tag type.
 ALTER STable stb_name DROP TAG tag_name;
 ```
 
-The tag will be removed automatically from all the subtables created using the super table as template once a tag is removed from a super table.
+The tag will be removed automatically from all the subtables, created using the super table as template, once a tag is removed from a super table.
 
 ### Change A Tag
 
@@ -102,7 +102,7 @@ The tag will be removed automatically from all the subtables created using the s
 ALTER STable stb_name CHANGE TAG old_tag_name new_tag_name;
 ```
 
-The tag name will be changed automatically for all the subtables created using the super table as template once a tag name is changed for a super table.
+The tag name will be changed automatically for all the subtables, created using the super table as template, once a tag name is changed for a super table.
 
 ### Change Tag Length
 
@@ -110,7 +110,7 @@ The tag name will be changed automatically for all the subtables created using t
 ALTER STable stb_name MODIFY TAG tag_name data_type(length);
 ```
 
-This command can be used to change (or increase, more specifically) the length of a tag of variable length types, like BINARY or NCHAR.
+This command can be used to change (or more specifically, increase) the length of a tag of variable length types, like BINARY or NCHAR.
 
 :::note
 Changing tag values can be applied to only subtables. All other tag operations, like add tag, remove tag, however, can be applied to only STable. If a new tag is added for a STable, the tag will be added with NULL value for all its subtables.

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

@@ -21,7 +21,7 @@ SELECT select_expr [, select_expr ...]
 
 ## Wildcard
 
-Wilcard \* can be used to specify all columns. The result includes only data columns for normal tables.
+Wildcard \* can be used to specify all columns. The result includes only data columns for normal tables.
 
 ```
 taos> SELECT * FROM d1001;
@@ -51,14 +51,14 @@ taos> SELECT * FROM meters;
 Query OK, 9 row(s) in set (0.002022s)
 ```
 
-Wildcard can be used with table name as prefix, both below SQL statements have same effects and return all columns.
+Wildcard can be used with table name as prefix. Both SQL statements below have the same effect and return all columns.
 
 ```SQL
 SELECT * FROM d1001;
 SELECT d1001.* FROM d1001;
 ```
 
-In JOIN query, however, with or without table name prefix will return different results. \* without table prefix will return all the columns of both tables, but \* with table name as prefix will return only the columns of that table.
+In a JOIN query, however, the results are different with or without a table name prefix. \* without table prefix will return all the columns of both tables, but \* with table name as prefix will return only the columns of that table.
 
 ```
 taos> SELECT * FROM d1001, d1003 WHERE d1001.ts=d1003.ts;
@@ -76,7 +76,7 @@ taos> SELECT d1001.* FROM d1001,d1003 WHERE d1001.ts = d1003.ts;
 Query OK, 1 row(s) in set (0.020443s)
 ```
 
-Wilcard \* can be used with some functions, but the result may be different depending on the function being used. For example, `count(*)` returns only one column, i.e. the number of rows; `first`, `last` and `last_row` return all columns of the selected row.
+Wildcard \* can be used with some functions, but the result may be different depending on the function being used. For example, `count(*)` returns only one column, i.e. the number of rows; `first`, `last` and `last_row` return all columns of the selected row.
 
 ```
 taos> SELECT COUNT(*) FROM d1001;
@@ -96,7 +96,7 @@ Query OK, 1 row(s) in set (0.000849s)
 
 ## Tags
 
-Starting from version 2.0.14, tag columns can be selected together with data columns when querying sub tables. Please note that, however, wildcard \* doesn't represent any tag column, that means tag columns must be specified explicitly like the example below.
+Starting from version 2.0.14, tag columns can be selected together with data columns when querying sub tables. Please note however, that, wildcard \* cannot be used to represent any tag column. This means that tag columns must be specified explicitly like the example below.
 
 ```
 taos> SELECT location, groupid, current FROM d1001 LIMIT 2;
@@ -109,7 +109,7 @@ Query OK, 2 row(s) in set (0.003112s)
 
 ## Get distinct values
 
-`DISTINCT` keyword can be used to get all the unique values of tag columns from a super table, it can also be used to get all the unique values of data columns from a table or subtable.
+`DISTINCT` keyword can be used to get all the unique values of tag columns from a super table. It can also be used to get all the unique values of data columns from a table or subtable.
 
 ```sql
 SELECT DISTINCT tag_name [, tag_name ...] FROM stb_name;
@@ -118,15 +118,15 @@ SELECT DISTINCT col_name [, col_name ...] FROM tb_name;
 
 :::info
 
-1. Configuration parameter `maxNumOfDistinctRes` in `taos.cfg` is used to control the number of rows to output. The minimum configurable value is 100,000, the maximum configurable value is 100,000,000, the default value is 1000,000. If the actual number of rows exceeds the value of this parameter, only the number of rows specified by this parameter will be output.
-2. It can't be guaranteed that the results selected by using `DISTINCT` on columns of `FLOAT` or `DOUBLE` are exactly unique because of the precision nature of floating numbers.
+1. Configuration parameter `maxNumOfDistinctRes` in `taos.cfg` is used to control the number of rows to output. The minimum configurable value is 100,000, the maximum configurable value is 100,000,000, the default value is 1,000,000. If the actual number of rows exceeds the value of this parameter, only the number of rows specified by this parameter will be output.
+2. It can't be guaranteed that the results selected by using `DISTINCT` on columns of `FLOAT` or `DOUBLE` are exactly unique because of the precision errors in floating point numbers.
 3. `DISTINCT` can't be used in the sub-query of a nested query statement, and can't be used together with aggregate functions, `GROUP BY` or `JOIN` in the same SQL statement.
 
 :::
 
 ## Columns Names of Result Set
 
-When using `SELECT`, the column names in the result set will be same as that in the select clause if `AS` is not used. `AS` can be used to rename the column names in the result set. For example
+When using `SELECT`, the column names in the result set will be the same as that in the select clause if `AS` is not used. `AS` can be used to rename the column names in the result set. For example
 
 ```
 taos> SELECT ts, ts AS primary_key_ts FROM d1001;
@@ -161,7 +161,7 @@ SELECT * FROM d1001;
 
 ## Special Query
 
-Some special query functionalities can be performed without `FORM` sub-clause. For example, below statement can be used to get the current database in use.
+Some special query functions can be invoked without `FROM` sub-clause. For example, the statement below can be used to get the current database in use.
 
 ```
 taos> SELECT DATABASE();
@@ -181,7 +181,7 @@ taos> SELECT DATABASE();
 Query OK, 1 row(s) in set (0.000184s)
 ```
 
-Below statement can be used to get the version of client or server.
+The statement below can be used to get the version of client or server.
 
 ```
 taos> SELECT CLIENT_VERSION();
@@ -197,7 +197,7 @@ taos> SELECT SERVER_VERSION();
 Query OK, 1 row(s) in set (0.000077s)
 ```
 
-Below statement is used to check the server status. One integer, like `1`, is returned if the server status is OK, otherwise an error code is returned. This is compatible with the status check for TDengine from connection pool or 3rd party tools, and can avoid the problem of losing the connection from a connection pool when using the wrong heartbeat checking SQL statement.
+The statement below is used to check the server status. An integer, like `1`, is returned if the server status is OK, otherwise an error code is returned. This is compatible with the status check for TDengine from connection pool or 3rd party tools, and can avoid the problem of losing the connection from a connection pool when using the wrong heartbeat checking SQL statement.
 
 ```
 taos> SELECT SERVER_STATUS();
@@ -284,7 +284,7 @@ taos> SELECT COUNT(tbname) FROM meters WHERE groupId > 2;
 Query OK, 1 row(s) in set (0.001091s)
 ```
 
-- Wildcard \* can be used to get all columns, or specific column names can be specified. Arithmetic operation can be performed on columns of number types, columns can be renamed in the result set.
+- Wildcard \* can be used to get all columns, or specific column names can be specified. Arithmetic operation can be performed on columns of numerical types, columns can be renamed in the result set.
 - Arithmetic operation on columns can't be used in where clause. For example, `where a*2>6;` is not allowed but `where a>6/2;` can be used instead for the same purpose.
 - Arithmetic operation on columns can't be used as the objectives of select statement. For example, `select min(2*a) from t;` is not allowed but `select 2*min(a) from t;` can be used instead.
 - Logical operation can be used in `WHERE` clause to filter numeric values, wildcard can be used to filter string values.
@@ -318,13 +318,13 @@ Logical operations in below table can be used in the `where` clause to filter th
 - Operator `like` is used together with wildcards to match strings
   - '%' matches 0 or any number of characters, '\_' matches any single ASCII character.
   - `\_` is used to match the \_ in the string.
-  - The maximum length of wildcard string is 100 bytes from version 2.1.6.1 (before that the maximum length is 20 bytes). `maxWildCardsLength` in `taos.cfg` can be used to control this threshold. Too long wildcard string may slowdown the execution performance of `LIKE` operator.
+  - The maximum length of wildcard string is 100 bytes from version 2.1.6.1 (before that the maximum length is 20 bytes). `maxWildCardsLength` in `taos.cfg` can be used to control this threshold. A very long wildcard string may slowdown the execution performance of `LIKE` operator.
 - `AND` keyword can be used to filter multiple columns simultaneously. AND/OR operation can be performed on single or multiple columns from version 2.3.0.0. However, before 2.3.0.0 `OR` can't be used on multiple columns.
 - For timestamp column, only one condition can be used; for other columns or tags, `OR` keyword can be used to combine multiple logical operators. For example, `((value > 20 AND value < 30) OR (value < 12))`.
   - From version 2.3.0.0, multiple conditions can be used on timestamp column, but the result set can only contain single time range.
 - From version 2.0.17.0, operator `BETWEEN AND` can be used in where clause, for example `WHERE col2 BETWEEN 1.5 AND 3.25` means the filter condition is equal to "1.5 ≤ col2 ≤ 3.25".
-- From version 2.1.4.0, operator `IN` can be used in the where clause. For example, `WHERE city IN ('California.SanFrancisco', 'California.SanDiego')`. For bool type, both `{true, false}` and `{0, 1}` are allowed, but integers other than 0 or 1 are not allowed. FLOAT and DOUBLE types are impacted by floating precision, only values that match the condition within the tolerance will be selected. Non-primary key column of timestamp type can be used with `IN`.
-- From version 2.3.0.0, regular expression is supported in the where clause with keyword `match` or `nmatch`, the regular expression is case insensitive.
+- From version 2.1.4.0, operator `IN` can be used in the where clause. For example, `WHERE city IN ('California.SanFrancisco', 'California.SanDiego')`. For bool type, both `{true, false}` and `{0, 1}` are allowed, but integers other than 0 or 1 are not allowed. FLOAT and DOUBLE types are impacted by floating point precision errors. Only values that match the condition within the tolerance will be selected. Non-primary key column of timestamp type can be used with `IN`.
+- From version 2.3.0.0, regular expression is supported in the where clause with keyword `match` or `nmatch`. The regular expression is case insensitive.
 
 ## Regular Expression
 
@@ -364,7 +364,7 @@ FROM temp_STable t1, temp_STable t2
 WHERE t1.ts = t2.ts AND t1.deviceid = t2.deviceid AND t1.status=0;
 ```
 
-Similary, join operation can be performed on the result set of multiple sub queries.
+Similarly, join operations can be performed on the result set of multiple sub queries.
 
 :::note
 Restrictions on join operation:
@@ -380,7 +380,7 @@ Restrictions on join operation:
 
 ## Nested Query
 
-Nested query is also called sub query, that means in a single SQL statement the result of inner query can be used as the data source of the outer query.
+Nested query is also called sub query. This means that in a single SQL statement the result of inner query can be used as the data source of the outer query.
 
 From 2.2.0.0, unassociated sub query can be used in the `FROM` clause. Unassociated means the sub query doesn't use the parameters in the parent query. More specifically, in the `tb_name_list` of `SELECT` statement, an independent SELECT statement can be used. So a complete nested query looks like:
 
@@ -390,14 +390,14 @@ SELECT ... FROM (SELECT ... FROM ...) ...;
 
 :::info
 
-- Only one layer of nesting is allowed, that means no sub query is allowed in a sub query
-- The result set returned by the inner query will be used as a "virtual table" by the outer query, the "virtual table" can be renamed using `AS` keyword for easy reference in the outer query.
+- Only one layer of nesting is allowed, that means no sub query is allowed within a sub query
+- The result set returned by the inner query will be used as a "virtual table" by the outer query. The "virtual table" can be renamed using `AS` keyword for easy reference in the outer query.
 - Sub query is not allowed in continuous query.
 - JOIN operation is allowed between tables/STables inside both inner and outer queries. Join operation can be performed on the result set of the inner query.
 - UNION operation is not allowed in either inner query or outer query.
-- The functionalities that can be used in the inner query is same as non-nested query.
-  - `ORDER BY` inside the inner query doesn't make any sense but will slow down the query performance significantly, so please avoid such usage.
-- Compared to the non-nested query, the functionalities that can be used in the outer query have such restrictions as:
+- The functions that can be used in the inner query are the same as those that can be used in a non-nested query.
+  - `ORDER BY` inside the inner query is unnecessary and will slow down the query performance significantly. It is best to avoid the use of `ORDER BY` inside the inner query.
+- Compared to the non-nested query, the functionality that can be used in the outer query has the following restrictions:
   - Functions
     - If the result set returned by the inner query doesn't contain timestamp column, then functions relying on timestamp can't be used in the outer query, like `TOP`, `BOTTOM`, `FIRST`, `LAST`, `DIFF`.
     - Functions that need to scan the data twice can't be used in the outer query, like `STDDEV`, `PERCENTILE`.
@@ -442,8 +442,8 @@ The sum of col1 and col2 for rows later than 2018-06-01 08:00:00.000 and whose c
 SELECT (col1 + col2) AS 'complex' FROM tb1 WHERE ts > '2018-06-01 08:00:00.000' AND col2 > 1.2 LIMIT 10 OFFSET 5;
 ```
 
-The rows in the past 10 minutes and whose col2 is bigger than 3.14 are selected and output to the result file `/home/testoutpu.csv` with below SQL statement:
+The rows in the past 10 minutes and whose col2 is bigger than 3.14 are selected and output to the result file `/home/testoutput.csv` with below SQL statement:
 
 ```SQL
-SELECT COUNT(*) FROM tb1 WHERE ts >= NOW - 10m AND col2 > 3.14 >> /home/testoutpu.csv;
+SELECT COUNT(*) FROM tb1 WHERE ts >= NOW - 10m AND col2 > 3.14 >> /home/testoutput.csv;
 ```

docs-en/12-taos-sql/index.md

@@ -7,8 +7,6 @@ This section explains the syntax of SQL to perform operations on databases, tabl
 
 TDengine SQL is the major interface for users to write data into or query from TDengine. For ease of use, the syntax is similar to that of standard SQL. However, please note that TDengine SQL is not standard SQL. For instance, TDengine doesn't provide a delete function for time series data and so corresponding statements are not provided in TDengine SQL.
 
-TDengine SQL doesn't support abbreviation for keywords. For example "SELECT" cannot be abbreviated as "SEL".
-
 Syntax Specifications used in this chapter:
 
 - The content inside <\> needs to be input by the user, excluding <\> itself.

docs-en/14-reference/07-tdinsight/index.md

@@ -61,7 +61,7 @@ sudo yum install \
 
 ## Automated deployment of TDinsight
 
-We provide an installation script [`TDinsight.sh`](https://github.com/taosdata/grafanaplugin/releases/latest/download/TDinsight.sh) script to allow users to configure the installation automatically and quickly.
+We provide an installation script [`TDinsight.sh`](https://github.com/taosdata/grafanaplugin/releases/latest/download/TDinsight.sh) to allow users to configure the installation automatically and quickly.
 
 You can download the script via `wget` or other tools:
 
@@ -300,7 +300,7 @@ This section contains the current information and status of the cluster, the ale
 
 ![TDengine Database TDinsight mnodes overview](./assets/TDinsight-3-mnodes.webp)
 
-1. **MNodes Status**: a simple table view of `show mnodes`. 2.
+1. **MNodes Status**: a simple table view of `show mnodes`.
 2. **MNodes Number**: similar to `DNodes Number`, the number of MNodes changes.
 
 ### Request
@@ -317,9 +317,9 @@ This section contains the current information and status of the cluster, the ale
 
 Database usage, repeated for each value of the variable `$database` i.e. multiple rows per database.
 
-1. **STables**: number of super tables. 2.
-2. **Total Tables**: number of all tables. 3.
-3. **Sub Tables**: the number of all super table sub-tables. 4.
+1. **STables**: number of super tables.
+2. **Total Tables**: number of all tables.
+3. **Sub Tables**: the number of all super table subtables.
 4. **Tables**: graph of all normal table numbers over time.
 5. **Tables Number Foreach VGroups**: The number of tables contained in each VGroups.
 
@@ -330,18 +330,18 @@ Database usage, repeated for each value of the variable `$database` i.e. multipl
 Data node resource usage display with repeated multiple rows for the variable `$fqdn` i.e., each data node. Includes.
 
 1. **Uptime**: the time elapsed since the dnode was created.
-2. **Has MNodes?**: whether the current dnode is a mnode. 3.
-3. **CPU Cores**: the number of CPU cores. 4.
-4. **VNodes Number**: the number of VNodes in the current dnode. 5.
-5. **VNodes Masters**: the number of vnodes in the master role. 6.
+2. **Has MNodes?**: whether the current dnode is a mnode.
+3. **CPU Cores**: the number of CPU cores.
+4. **VNodes Number**: the number of VNodes in the current dnode.
+5. **VNodes Masters**: the number of vnodes in the master role.
 6. **Current CPU Usage of taosd**: CPU usage rate of taosd processes.
 7. **Current Memory Usage of taosd**: memory usage of taosd processes.
 8. **Disk Used**: The total disk usage percentage of the taosd data directory.
-9. **CPU Usage**: Process and system CPU usage. 10.
+9. **CPU Usage**: Process and system CPU usage.
 10. **RAM Usage**: Time series view of RAM usage metrics.
 11. **Disk Used**: Disks used at each level of multi-level storage (default is level0).
 12. **Disk Increasing Rate per Minute**: Percentage increase or decrease in disk usage per minute.
-13. **Disk IO**: Disk IO rate. 14.
+13. **Disk IO**: Disk IO rate.
 14. **Net IO**: Network IO, the aggregate network IO rate in addition to the local network.
 
 ### Login History
@@ -376,7 +376,7 @@ TDinsight installed via the `TDinsight.sh` script can be cleaned up using the co
 To completely uninstall TDinsight during a manual installation, you need to clean up the following.
 
 1. the TDinsight Dashboard in Grafana.
-2. the Data Source in Grafana. 3.
+2. the Data Source in Grafana.
 3. remove the `tdengine-datasource` plugin from the plugin installation directory.
 
 ## Integrated Docker Example

docs-en/20-third-party/01-grafana.mdx

@@ -23,7 +23,7 @@ You can download The Grafana plugin for TDengine from <https://github.com/taosda
 
 Recommend using the [``grafana-cli`` command-line tool](https://grafana.com/docs/grafana/latest/administration/cli/) for plugin installation.
 
-``bash
+```bash
 sudo -u grafana grafana-cli \
   --pluginUrl https://github.com/taosdata/grafanaplugin/releases/download/v3.1.4/tdengine-datasource-3.1.4.zip \
   plugins install tdengine-datasource
@@ -88,7 +88,7 @@ Go back to the main interface to create the Dashboard, click Add Query to enter
 
 As shown above, select the `TDengine` data source in the `Query` and enter the corresponding SQL in the query box below for query.
 
-- INPUT SQL: enter the statement to be queried (the result set of the SQL statement should be two columns and multiple rows), for example: `select avg(mem_system) from log.dn where ts >= $from and ts < $to interval($interval)`, where, from, to and interval are built-in variables of the TDengine plugin, indicating the range and time interval of queries fetched from the Grafana plugin panel. In addition to the built-in variables, ` custom template variables are also supported.
+- INPUT SQL: enter the statement to be queried (the result set of the SQL statement should be two columns and multiple rows), for example: `select avg(mem_system) from log.dn where ts >= $from and ts < $to interval($interval)`, where, from, to and interval are built-in variables of the TDengine plugin, indicating the range and time interval of queries fetched from the Grafana plugin panel. In addition to the built-in variables, custom template variables are also supported.
 - ALIAS BY: This allows you to set the current query alias.
 - GENERATE SQL: Clicking this button will automatically replace the corresponding variables and generate the final executed statement.
 

docs-en/20-third-party/09-emq-broker.md

@@ -3,7 +3,7 @@ sidebar_label: EMQX Broker
 title: EMQX Broker writing
 ---
 
-MQTT is a popular IoT data transfer protocol, [EMQX](https://github.com/emqx/emqx) is an open-source MQTT Broker software, without any code, only need to use "rules" in EMQX Dashboard to do simple configuration. You can write MQTT data directly to TDengine. EMQX supports saving data to TDengine by sending it to web services and provides a native TDengine driver for direct saving in the Enterprise Edition. Please refer to the [EMQX official documentation](https://www.emqx.io/docs/en/v4.4/rule/rule-engine.html) for details on how to use it. tdengine).
+MQTT is a popular IoT data transfer protocol, [EMQX](https://github.com/emqx/emqx) is an open-source MQTT Broker software, you can write MQTT data directly to TDengine without any code, you only need to use "rules" in EMQX Dashboard to create a simple configuration. EMQX supports saving data to TDengine by sending it to web services and provides a native TDengine driver for direct saving in the Enterprise Edition. Please refer to the [EMQX official documentation](https://www.emqx.io/docs/en/v4.4/rule/rule-engine.html) for details on how to use it.).
 
 ## Prerequisites
 

docs-en/20-third-party/11-kafka.md

@@ -228,7 +228,7 @@ taos> select * from meters;
 Query OK, 4 row(s) in set (0.004208s)
 ```
 
-If you see the above data, the synchronization is successful. If not, check the logs of Kafka Connect. For detailed description of configuration parameters, see [Configuration Reference](#Configuration Reference).
+If you see the above data, the synchronization is successful. If not, check the logs of Kafka Connect. For detailed description of configuration parameters, see [Configuration Reference](#configuration-reference).
 
 ## The use of TDengine Source Connector
 

docs-examples/c/async_query_example.c

@@ -182,14 +182,14 @@ int main() {
 // query callback ...
 // ts current voltage phase location groupid
 // numOfRow = 8
-// 1538548685000 10.300000 219 0.310000 beijing.chaoyang 2
-// 1538548695000 12.600000 218 0.330000 beijing.chaoyang 2
-// 1538548696800 12.300000 221 0.310000 beijing.chaoyang 2
-// 1538548696650 10.300000 218 0.250000 beijing.chaoyang 3
-// 1538548685500 11.800000 221 0.280000 beijing.haidian 2
-// 1538548696600 13.400000 223 0.290000 beijing.haidian 2
-// 1538548685000 10.800000 223 0.290000 beijing.haidian 3
-// 1538548686500 11.500000 221 0.350000 beijing.haidian 3
+// 1538548685500 11.800000 221 0.280000 california.losangeles 2
+// 1538548696600 13.400000 223 0.290000 california.losangeles 2
+// 1538548685000 10.800000 223 0.290000 california.losangeles 3
+// 1538548686500 11.500000 221 0.350000 california.losangeles 3
+// 1538548685000 10.300000 219 0.310000 california.sanfrancisco 2
+// 1538548695000 12.600000 218 0.330000 california.sanfrancisco 2
+// 1538548696800 12.300000 221 0.310000 california.sanfrancisco 2
+// 1538548696650 10.300000 218 0.250000 california.sanfrancisco 3
 // numOfRow = 0
 // no more data, close the connection.
 // ANCHOR_END: demo
\ No newline at end of file

docs-examples/csharp/AsyncQueryExample.cs

@@ -224,15 +224,15 @@ namespace TDengineExample
 }
 
 //output:
-//Connect to TDengine success
-//8 rows async retrieved
-
-//1538548685000   |       10.3    |       219     |       0.31    |       beijing.chaoyang        |       2       |
-//1538548695000   |       12.6    |       218     |       0.33    |       beijing.chaoyang        |       2       |
-//1538548696800   |       12.3    |       221     |       0.31    |       beijing.chaoyang        |       2       |
-//1538548696650   |       10.3    |       218     |       0.25    |       beijing.chaoyang        |       3       |
-//1538548685500   |       11.8    |       221     |       0.28    |       beijing.haidian         |       2       |
-//1538548696600   |       13.4    |       223     |       0.29    |       beijing.haidian         |       2       |
-//1538548685000   |       10.8    |       223     |       0.29    |       beijing.haidian         |       3       |
-//1538548686500   |       11.5    |       221     |       0.35    |       beijing.haidian         |       3       |
-//async retrieve complete.
+// Connect to TDengine success
+// 8 rows async retrieved
+
+// 1538548685500   |       11.8    |       221     |       0.28    |       california.losangeles   |       2       |
+// 1538548696600   |       13.4    |       223     |       0.29    |       california.losangeles   |       2       |
+// 1538548685000   |       10.8    |       223     |       0.29    |       california.losangeles   |       3       |
+// 1538548686500   |       11.5    |       221     |       0.35    |       california.losangeles   |       3       |
+// 1538548685000   |       10.3    |       219     |       0.31    |       california.sanfrancisco         |       2       |
+// 1538548695000   |       12.6    |       218     |       0.33    |       california.sanfrancisco         |       2       |
+// 1538548696800   |       12.3    |       221     |       0.31    |       california.sanfrancisco         |       2       |
+// 1538548696650   |       10.3    |       218     |       0.25    |       california.sanfrancisco         |       3       |
+// async retrieve complete.
\ No newline at end of file

docs-examples/python/conn_native_pandas.py

@@ -13,7 +13,7 @@ print(df.head(3))
 # output:
 # RangeIndex(start=0, stop=8, step=1)
 # <class 'pandas._libs.tslibs.timestamps.Timestamp'>
-#                        ts  current  voltage  phase          location  groupid
-# 0 2018-10-03 14:38:05.000     10.3      219   0.31  beijing.chaoyang        2
-# 1 2018-10-03 14:38:15.000     12.6      218   0.33  beijing.chaoyang        2
-# 2 2018-10-03 14:38:16.800     12.3      221   0.31  beijing.chaoyang        2
+#                        ts  current  ...               location  groupid
+# 0 2018-10-03 14:38:05.500     11.8  ...  california.losangeles        2
+# 1 2018-10-03 14:38:16.600     13.4  ...  california.losangeles        2
+# 2 2018-10-03 14:38:05.000     10.8  ...  california.losangeles        3

docs-examples/python/conn_rest_pandas.py

@@ -11,9 +11,9 @@ print(type(df.ts[0]))
 print(df.head(3))
 
 # output:
-# <class 'datetime.datetime'>
 # RangeIndex(start=0, stop=8, step=1)
-#                                  ts  current  ...          location  groupid
-# 0         2018-10-03 14:38:05+08:00     10.3  ...  beijing.chaoyang        2
-# 1         2018-10-03 14:38:15+08:00     12.6  ...  beijing.chaoyang        2
-# 2  2018-10-03 14:38:16.800000+08:00     12.3  ...  beijing.chaoyang        2
+# <class 'pandas._libs.tslibs.timestamps.Timestamp'>
+#                                 ts  current  ...               location  groupid
+# 0 2018-10-03 06:38:05.500000+00:00     11.8  ...  california.losangeles        2
+# 1 2018-10-03 06:38:16.600000+00:00     13.4  ...  california.losangeles        2
+# 2        2018-10-03 06:38:05+00:00     10.8  ...  california.losangeles        3

docs-examples/python/connect_rest_examples.py

@@ -38,8 +38,7 @@ for row in data:
 # inserted row count: 8
 # queried row count: 3
 # ['ts', 'current', 'voltage', 'phase', 'location', 'groupid']
-# [datetime.datetime(2018, 10, 3, 14, 38, 5, tzinfo=datetime.timezone(datetime.timedelta(seconds=28800), '+08:00')), 10.3, 219, 0.31, 'beijing.chaoyang', 2]
-# [datetime.datetime(2018, 10, 3, 14, 38, 15, tzinfo=datetime.timezone(datetime.timedelta(seconds=28800), '+08:00')), 12.6, 218, 0.33, 'beijing.chaoyang', 2]
-# [datetime.datetime(2018, 10, 3, 14, 38, 16, 800000, tzinfo=datetime.timezone(datetime.timedelta(seconds=28800), '+08:00')), 12.3, 221, 0.31, 'beijing.chaoyang', 2]
-
+# [datetime.datetime(2018, 10, 3, 14, 38, 5, 500000, tzinfo=datetime.timezone(datetime.timedelta(seconds=28800), '+08:00')), 11.8, 221, 0.28, 'california.losangeles', 2]
+# [datetime.datetime(2018, 10, 3, 14, 38, 16, 600000, tzinfo=datetime.timezone(datetime.timedelta(seconds=28800), '+08:00')), 13.4, 223, 0.29, 'california.losangeles', 2]
+# [datetime.datetime(2018, 10, 3, 14, 38, 5, tzinfo=datetime.timezone(datetime.timedelta(seconds=28800), '+08:00')), 10.8, 223, 0.29, 'california.losangeles', 3]
 # ANCHOR_END: basic

docs-examples/python/json_protocol_example.py

@@ -5,9 +5,9 @@ from taos import SmlProtocol, SmlPrecision
 
 lines = [{"metric": "meters.current", "timestamp": 1648432611249, "value": 10.3, "tags": {"location": "California.SanFrancisco", "groupid": 2}},
          {"metric": "meters.voltage", "timestamp": 1648432611249, "value": 219,
-             "tags": {"location": "California.LosAngeles", "groupid": 1}},
+          "tags": {"location": "California.LosAngeles", "groupid": 1}},
          {"metric": "meters.current", "timestamp": 1648432611250, "value": 12.6,
-             "tags": {"location": "California.SanFrancisco", "groupid": 2}},
+          "tags": {"location": "California.SanFrancisco", "groupid": 2}},
          {"metric": "meters.voltage", "timestamp": 1648432611250, "value": 221, "tags": {"location": "California.LosAngeles", "groupid": 1}}]
 
 

docs-examples/python/query_example.py

@@ -12,10 +12,10 @@ def query_api_demo(conn: taos.TaosConnection):
 
 
 # field count: 7
-# meta of files[1]: {name: ts, type: 9, bytes: 8}
+# meta of fields[1]: {name: ts, type: 9, bytes: 8}
 # ======================Iterate on result=========================
-# ('d1001', datetime.datetime(2018, 10, 3, 14, 38, 5), 10.300000190734863, 219, 0.3100000023841858, 'California.SanFrancisco', 2)
-# ('d1001', datetime.datetime(2018, 10, 3, 14, 38, 15), 12.600000381469727, 218, 0.33000001311302185, 'California.SanFrancisco', 2)
+# ('d1003', datetime.datetime(2018, 10, 3, 14, 38, 5, 500000), 11.800000190734863, 221, 0.2800000011920929, 'california.losangeles', 2)
+# ('d1003', datetime.datetime(2018, 10, 3, 14, 38, 16, 600000), 13.399999618530273, 223, 0.28999999165534973, 'california.losangeles', 2)
 # ANCHOR_END: iter
 
 # ANCHOR: fetch_all
@@ -29,8 +29,8 @@ def fetch_all_demo(conn: taos.TaosConnection):
 
 # row count: 2
 # ===============all data===================
-# [{'ts': datetime.datetime(2018, 10, 3, 14, 38, 5), 'current': 10.300000190734863},
-# {'ts': datetime.datetime(2018, 10, 3, 14, 38, 15), 'current': 12.600000381469727}]
+# [{'ts': datetime.datetime(2018, 10, 3, 14, 38, 5, 500000), 'current': 11.800000190734863},
+# {'ts': datetime.datetime(2018, 10, 3, 14, 38, 16, 600000), 'current': 13.399999618530273}]
 # ANCHOR_END: fetch_all
 
 if __name__ == '__main__':

include/common/tcommon.h

@@ -105,12 +105,14 @@ typedef struct SColumnInfoData {
 } SColumnInfoData;
 
 typedef struct SQueryTableDataCond {
-  STimeWindow  twindow;
+  //STimeWindow  twindow;
   int32_t      order;  // desc|asc order to iterate the data block
   int32_t      numOfCols;
   SColumnInfo *colList;
   bool         loadExternalRows;  // load external rows or not
   int32_t      type;              // data block load type:
+  int32_t      numOfTWindows;
+  STimeWindow *twindows;
 } SQueryTableDataCond;
 
 void*   blockDataDestroy(SSDataBlock* pBlock);

include/dnode/mnode/mnode.h

@@ -81,7 +81,7 @@ int32_t mndGetLoad(SMnode *pMnode, SMnodeLoad *pLoad);
  * @param pMsg The request msg.
  * @return int32_t 0 for success, -1 for failure.
  */
-int32_t mndProcessMsg(SRpcMsg *pMsg);
+int32_t mndProcessRpcMsg(SRpcMsg *pMsg);
 int32_t mndProcessSyncMsg(SRpcMsg *pMsg);
 
 /**

include/libs/catalog/catalog.h

@@ -52,23 +52,31 @@ typedef struct SUserAuthInfo {
   AUTH_TYPE type;
 } SUserAuthInfo;
 
+typedef struct SDbInfo {
+  int32_t vgVer;
+  int32_t tbNum;
+  int64_t dbId;
+} SDbInfo;
+
 typedef struct SCatalogReq {
-  SArray *pTableMeta;     // element is SNAME
   SArray *pDbVgroup;      // element is db full name
+  SArray *pDbCfg;         // element is db full name
+  SArray *pDbInfo;        // element is db full name
+  SArray *pTableMeta;     // element is SNAME
   SArray *pTableHash;     // element is SNAME
   SArray *pUdf;           // element is udf name
-  SArray *pDbCfg;         // element is db full name
   SArray *pIndex;         // element is index name
   SArray *pUser;          // element is SUserAuthInfo
   bool    qNodeRequired;  // valid qnode
 } SCatalogReq;
 
 typedef struct SMetaData {
-  SArray    *pTableMeta;  // SArray<STableMeta*>
   SArray    *pDbVgroup;   // SArray<SArray<SVgroupInfo>*>
+  SArray    *pDbCfg;      // SArray<SDbCfgInfo>
+  SArray    *pDbInfo;     // SArray<SDbInfo>
+  SArray    *pTableMeta;  // SArray<STableMeta*>
   SArray    *pTableHash;  // SArray<SVgroupInfo>
   SArray    *pUdfList;    // SArray<SFuncInfo>
-  SArray    *pDbCfg;      // SArray<SDbCfgInfo>
   SArray    *pIndex;      // SArray<SIndexInfo>
   SArray    *pUser;       // SArray<bool>
   SArray    *pQnodeList;  // SArray<SQueryNodeAddr>
@@ -269,6 +277,8 @@ int32_t catalogChkAuth(SCatalog* pCtg, void *pRpc, const SEpSet* pMgmtEps, const
 
 int32_t catalogUpdateUserAuthInfo(SCatalog* pCtg, SGetUserAuthRsp* pAuth);
 
+int32_t catalogUpdateVgEpSet(SCatalog* pCtg, const char* dbFName, int32_t vgId, SEpSet *epSet);
+
 
 int32_t ctgdLaunchAsyncCall(SCatalog* pCtg, void *pTrans, const SEpSet* pMgmtEps, uint64_t reqId);
 

include/libs/index/index.h

@@ -192,11 +192,16 @@ void        indexTermDestroy(SIndexTerm* p);
 void indexInit();
 
 /* index filter */
+typedef struct SIndexMetaArg {
+  void*    metaHandle;
+  uint64_t suid;
+} SIndexMetaArg;
+
 typedef enum { SFLT_NOT_INDEX, SFLT_COARSE_INDEX, SFLT_ACCURATE_INDEX } SIdxFltStatus;
 
 SIdxFltStatus idxGetFltStatus(SNode* pFilterNode);
 
-int32_t doFilterTag(const SNode* pFilterNode, SArray* result);
+int32_t doFilterTag(const SNode* pFilterNode, SIndexMetaArg* metaArg, SArray* result);
 /*
  * destory index env
  *

include/libs/nodes/plannodes.h

@@ -56,6 +56,9 @@ typedef struct SScanLogicNode {
   int8_t             intervalUnit;
   int8_t             slidingUnit;
   SNode*             pTagCond;
+  int8_t             triggerType;
+  int64_t            watermark;
+  int16_t            tsColId;
 } SScanLogicNode;
 
 typedef struct SJoinLogicNode {
@@ -216,6 +219,9 @@ typedef struct STableScanPhysiNode {
   int64_t        sliding;
   int8_t         intervalUnit;
   int8_t         slidingUnit;
+  int8_t         triggerType;
+  int64_t        watermark;
+  int16_t        tsColId;
 } STableScanPhysiNode;
 
 typedef STableScanPhysiNode STableSeqScanPhysiNode;

include/libs/qcom/query.h

@@ -43,6 +43,12 @@ typedef enum {
   TASK_TYPE_TEMP,
 } ETaskType;
 
+typedef enum {
+  TARGET_TYPE_MNODE = 1,
+  TARGET_TYPE_VNODE,
+  TARGET_TYPE_OTHER,
+} ETargetType;
+
 typedef struct STableComInfo {
   uint8_t  numOfTags;     // the number of tags in schema
   uint8_t  precision;     // the number of precision
@@ -126,11 +132,18 @@ typedef struct SDataBuf {
   void*    handle;
 } SDataBuf;
 
+typedef struct STargetInfo {
+  ETargetType type;
+  char        dbFName[TSDB_DB_FNAME_LEN]; // used to update db's vgroup epset
+  int32_t     vgId;
+} STargetInfo;
+
 typedef int32_t (*__async_send_cb_fn_t)(void* param, const SDataBuf* pMsg, int32_t code);
 typedef int32_t (*__async_exec_fn_t)(void* param);
 
 typedef struct SMsgSendInfo {
   __async_send_cb_fn_t fp;  // async callback function
+  STargetInfo          target; // for update epset
   void*                param;
   uint64_t             requestId;
   uint64_t             requestObjRefId;

source/client/src/clientImpl.c

@@ -729,23 +729,55 @@ static void destroySendMsgInfo(SMsgSendInfo* pMsgBody) {
   taosMemoryFreeClear(pMsgBody);
 }
 
+void updateTargetEpSet(SMsgSendInfo* pSendInfo, STscObj* pTscObj, SRpcMsg* pMsg, SEpSet* pEpSet) {
+  if (NULL == pEpSet) {
+    return;
+  }
+  
+  switch (pSendInfo->target.type) {
+    case TARGET_TYPE_MNODE:
+      if (NULL == pTscObj) {
+        tscError("mnode epset changed but not able to update it, reqObjRefId:%" PRIx64, pSendInfo->requestObjRefId);
+        return;
+      }
+
+      updateEpSet_s(&pTscObj->pAppInfo->mgmtEp, pEpSet);   
+      break;
+    case TARGET_TYPE_VNODE: {
+      if (NULL == pTscObj) {
+        tscError("vnode epset changed but not able to update it, reqObjRefId:%" PRIx64, pSendInfo->requestObjRefId);
+        return;
+      }
+
+      SCatalog* pCatalog = NULL;
+      int32_t code = catalogGetHandle(pTscObj->pAppInfo->clusterId, &pCatalog);
+      if (code != TSDB_CODE_SUCCESS) {
+        tscError("fail to get catalog handle, clusterId:%" PRIx64 ", error %s", pTscObj->pAppInfo->clusterId, tstrerror(code));
+        return;
+      }
+    
+      catalogUpdateVgEpSet(pCatalog, pSendInfo->target.dbFName, pSendInfo->target.vgId, pEpSet);
+      break;
+    }
+    default:
+      tscDebug("epset changed, not updated, msgType %s", TMSG_INFO(pMsg->msgType));
+      break;
+  }
+}
+
+
 void processMsgFromServer(void* parent, SRpcMsg* pMsg, SEpSet* pEpSet) {
   SMsgSendInfo* pSendInfo = (SMsgSendInfo*)pMsg->info.ahandle;
   assert(pMsg->info.ahandle != NULL);
+  SRequestObj* pRequest = NULL;
+  STscObj* pTscObj = NULL;
 
   if (pSendInfo->requestObjRefId != 0) {
     SRequestObj* pRequest = (SRequestObj*)taosAcquireRef(clientReqRefPool, pSendInfo->requestObjRefId);
     assert(pRequest->self == pSendInfo->requestObjRefId);
 
     pRequest->metric.rsp = taosGetTimestampUs();
-
-    //STscObj* pTscObj = pRequest->pTscObj;
-    //if (pEpSet) {
-    //  if (!isEpsetEqual(&pTscObj->pAppInfo->mgmtEp.epSet, pEpSet)) {
-    //    updateEpSet_s(&pTscObj->pAppInfo->mgmtEp, pEpSet);
-    //  }
-    //}
-
+    pTscObj = pRequest->pTscObj;
     /*
      * There is not response callback function for submit response.
      * The actual inserted number of points is the first number.
@@ -762,6 +794,8 @@ void processMsgFromServer(void* parent, SRpcMsg* pMsg, SEpSet* pEpSet) {
     taosReleaseRef(clientReqRefPool, pSendInfo->requestObjRefId);
   }
 
+  updateTargetEpSet(pSendInfo, pTscObj, pMsg, pEpSet);
+
   SDataBuf buf = {.len = pMsg->contLen, .pData = NULL, .handle = pMsg->info.handle};
 
   if (pMsg->contLen > 0) {

source/common/src/tdatablock.c

@@ -1153,7 +1153,9 @@ void colInfoDataCleanup(SColumnInfoData* pColumn, uint32_t numOfRows) {
   if (IS_VAR_DATA_TYPE(pColumn->info.type)) {
     pColumn->varmeta.length = 0;
   } else {
-    memset(pColumn->nullbitmap, 0, BitmapLen(numOfRows));
+    if (pColumn->nullbitmap != NULL) {
+      memset(pColumn->nullbitmap, 0, BitmapLen(numOfRows));
+    }
   }
 }
 

source/common/src/tglobal.c

@@ -293,7 +293,7 @@ int32_t taosAddClientLogCfg(SConfig *pCfg) {
   if (cfgAddInt32(pCfg, "jniDebugFlag", jniDebugFlag, 0, 255, 1) != 0) return -1;
   if (cfgAddInt32(pCfg, "simDebugFlag", 143, 0, 255, 1) != 0) return -1;
   if (cfgAddInt32(pCfg, "debugFlag", 0, 0, 255, 1) != 0) return -1;
-  if (cfgAddInt32(pCfg, "idxDebugFlag", 0, 0, 255, 1) != 0) return -1;
+  if (cfgAddInt32(pCfg, "idxDebugFlag", idxDebugFlag, 0, 255, 1) != 0) return -1;
   return 0;
 }
 

source/common/src/trow.c

@@ -1191,9 +1191,9 @@ bool tdGetTpRowDataOfCol(STSRowIter *pIter, col_type_t colType, int32_t offset,
 }
 
 static FORCE_INLINE int32_t compareKvRowColId(const void *key1, const void *key2) {
-  if (*(int16_t *)key1 > ((SColIdx *)key2)->colId) {
+  if (*(col_id_t *)key1 > ((SKvRowIdx *)key2)->colId) {
     return 1;
-  } else if (*(int16_t *)key1 < ((SColIdx *)key2)->colId) {
+  } else if (*(col_id_t *)key1 < ((SKvRowIdx *)key2)->colId) {
     return -1;
   } else {
     return 0;

source/common/src/ttime.c

@@ -528,14 +528,14 @@ int32_t convertStringToTimestamp(int16_t type, char *inputData, int64_t timePrec
     }
     taosMemoryFree(newColData);
   } else if (type == TSDB_DATA_TYPE_NCHAR) {
-    newColData = taosMemoryCalloc(1,  charLen / TSDB_NCHAR_SIZE + 1);
+    newColData = taosMemoryCalloc(1,  charLen + TSDB_NCHAR_SIZE);
     int len = taosUcs4ToMbs((TdUcs4 *)varDataVal(inputData), charLen, newColData);
     if (len < 0){
       taosMemoryFree(newColData);
       return TSDB_CODE_FAILED;
     }
     newColData[len] = 0;
-    bool ret = taosParseTime(newColData, timeVal, len + 1, (int32_t)timePrec, tsDaylight);
+    int32_t ret = taosParseTime(newColData, timeVal, len + 1, (int32_t)timePrec, tsDaylight);
     if (ret != TSDB_CODE_SUCCESS) {
       taosMemoryFree(newColData);
       return ret;

source/dnode/mgmt/mgmt_mnode/src/mmWorker.c

@@ -40,7 +40,7 @@ static void mmProcessQueue(SQueueInfo *pInfo, SRpcMsg *pMsg) {
       break;
     default:
       pMsg->info.node = pMgmt->pMnode;
-      code = mndProcessMsg(pMsg);
+      code = mndProcessRpcMsg(pMsg);
   }
 
   if (IsReq(pMsg) && pMsg->info.handle != NULL && code != TSDB_CODE_ACTION_IN_PROGRESS) {

source/dnode/mnode/impl/inc/mndInt.h

@@ -75,13 +75,12 @@ typedef struct {
 } STelemMgmt;
 
 typedef struct {
-  SWal      *pWal;
-  sem_t      syncSem;
-  int64_t    sync;
-  bool       standby;
-  bool       restored;
-  int32_t    errCode;
-  int32_t    transId;
+  SWal   *pWal;
+  sem_t   syncSem;
+  int64_t sync;
+  bool    standby;
+  int32_t errCode;
+  int32_t transId;
 } SSyncMgmt;
 
 typedef struct {
@@ -90,34 +89,45 @@ typedef struct {
 } SGrantInfo;
 
 typedef struct SMnode {
-  int32_t       selfDnodeId;
-  int64_t       clusterId;
-  TdThread      thread;
-  bool          deploy;
-  bool          stopped;
-  int8_t        replica;
-  int8_t        selfIndex;
-  SReplica      replicas[TSDB_MAX_REPLICA];
-  char         *path;
-  int64_t       checkTime;
-  SSdb         *pSdb;
-  SMgmtWrapper *pWrapper;
-  SArray       *pSteps;
-  SQHandle     *pQuery;
-  SShowMgmt     showMgmt;
-  SProfileMgmt  profileMgmt;
-  STelemMgmt    telemMgmt;
-  SSyncMgmt     syncMgmt;
-  SHashObj     *infosMeta;
-  SHashObj     *perfsMeta;
-  SGrantInfo    grant;
-  MndMsgFp      msgFp[TDMT_MAX];
-  SMsgCb        msgCb;
+  int32_t        selfDnodeId;
+  int64_t        clusterId;
+  TdThread       thread;
+  TdThreadRwlock lock;
+  int32_t        rpcRef;
+  int32_t        syncRef;
+  bool           stopped;
+  bool           restored;
+  bool           deploy;
+  int8_t         replica;
+  int8_t         selfIndex;
+  SReplica       replicas[TSDB_MAX_REPLICA];
+  char          *path;
+  int64_t        checkTime;
+  SSdb          *pSdb;
+  SArray        *pSteps;
+  SQHandle      *pQuery;
+  SHashObj      *infosMeta;
+  SHashObj      *perfsMeta;
+  SShowMgmt      showMgmt;
+  SProfileMgmt   profileMgmt;
+  STelemMgmt     telemMgmt;
+  SSyncMgmt      syncMgmt;
+  SGrantInfo     grant;
+  MndMsgFp       msgFp[TDMT_MAX];
+  SMsgCb         msgCb;
 } SMnode;
 
 void    mndSetMsgHandle(SMnode *pMnode, tmsg_t msgType, MndMsgFp fp);
 int64_t mndGenerateUid(char *name, int32_t len);
 
+int32_t mndAcquireRpcRef(SMnode *pMnode);
+void    mndReleaseRpcRef(SMnode *pMnode);
+void    mndSetRestore(SMnode *pMnode, bool restored);
+void    mndSetStop(SMnode *pMnode);
+bool    mndGetStop(SMnode *pMnode);
+int32_t mndAcquireSyncRef(SMnode *pMnode);
+void    mndReleaseSyncRef(SMnode *pMnode);
+
 #ifdef __cplusplus
 }
 #endif

source/dnode/mnode/impl/src/mnode.c → source/dnode/mnode/impl/src/mndMain.c

@@ -85,7 +85,7 @@ static void *mndThreadFp(void *param) {
   while (1) {
     lastTime++;
     taosMsleep(100);
-    if (pMnode->stopped) break;
+    if (mndGetStop(pMnode)) break;
 
     if (lastTime % (tsTransPullupInterval * 10) == 0) {
       mndPullupTrans(pMnode);
@@ -118,7 +118,6 @@ static int32_t mndInitTimer(SMnode *pMnode) {
 }
 
 static void mndCleanupTimer(SMnode *pMnode) {
-  pMnode->stopped = true;
   if (taosCheckPthreadValid(pMnode->thread)) {
     taosThreadJoin(pMnode->thread, NULL);
     taosThreadClear(&pMnode->thread);
@@ -335,15 +334,19 @@ void mndClose(SMnode *pMnode) {
 int32_t mndStart(SMnode *pMnode) {
   mndSyncStart(pMnode);
   if (pMnode->deploy) {
-    if (sdbDeploy(pMnode->pSdb) != 0) return -1;
-    pMnode->syncMgmt.restored = true;
+    if (sdbDeploy(pMnode->pSdb) != 0) {
+      mError("failed to deploy sdb while start mnode");
+      return -1;
+    }
+    mndSetRestore(pMnode, true);
   }
   return mndInitTimer(pMnode);
 }
 
 void mndStop(SMnode *pMnode) {
+  mndSetStop(pMnode);
   mndSyncStop(pMnode);
-  return mndCleanupTimer(pMnode);
+  mndCleanupTimer(pMnode);
 }
 
 int32_t mndProcessSyncMsg(SRpcMsg *pMsg) {
@@ -362,6 +365,11 @@ int32_t mndProcessSyncMsg(SRpcMsg *pMsg) {
     return TAOS_SYNC_PROPOSE_OTHER_ERROR;
   }
 
+  if (mndAcquireSyncRef(pMnode) != 0) {
+    mError("failed to process sync msg:%p type:%s since %s", pMsg, TMSG_INFO(pMsg->msgType), terrstr());
+    return TAOS_SYNC_PROPOSE_OTHER_ERROR;
+  }
+
   char  logBuf[512];
   char *syncNodeStr = sync2SimpleStr(pMgmt->sync);
   snprintf(logBuf, sizeof(logBuf), "==vnodeProcessSyncReq== msgType:%d, syncNode: %s", pMsg->msgType, syncNodeStr);
@@ -405,59 +413,45 @@ int32_t mndProcessSyncMsg(SRpcMsg *pMsg) {
     code = TAOS_SYNC_PROPOSE_OTHER_ERROR;
   }
 
+  mndReleaseSyncRef(pMnode);
   return code;
 }
 
-static int32_t mndCheckMnodeMaster(SRpcMsg *pMsg) {
-  if (!IsReq(pMsg)) return 0;
-  if (mndIsMaster(pMsg->info.node)) return 0;
+static int32_t mndCheckMnodeState(SRpcMsg *pMsg) {
+  if (mndAcquireRpcRef(pMsg->info.node) == 0) return 0;
 
-  if (pMsg->msgType == TDMT_MND_MQ_TIMER || pMsg->msgType == TDMT_MND_TELEM_TIMER ||
-      pMsg->msgType == TDMT_MND_TRANS_TIMER) {
-    return -1;
-  }
-  mError("msg:%p, failed to check master since %s, app:%p type:%s", pMsg, terrstr(), pMsg->info.ahandle,
-         TMSG_INFO(pMsg->msgType));
+  if (IsReq(pMsg) && pMsg->msgType != TDMT_MND_MQ_TIMER && pMsg->msgType != TDMT_MND_TELEM_TIMER &&
+      pMsg->msgType != TDMT_MND_TRANS_TIMER) {
+    mError("msg:%p, failed to check mnode state since %s, app:%p type:%s", pMsg, terrstr(), pMsg->info.ahandle,
+           TMSG_INFO(pMsg->msgType));
 
-  SEpSet epSet = {0};
-  mndGetMnodeEpSet(pMsg->info.node, &epSet);
+    SEpSet epSet = {0};
+    mndGetMnodeEpSet(pMsg->info.node, &epSet);
 
-#if 0
-  mTrace("msg:%p, is redirected, num:%d use:%d", pMsg, epSet.numOfEps, epSet.inUse);
-  for (int32_t i = 0; i < epSet.numOfEps; ++i) {
-    mTrace("mnode index:%d %s:%u", i, epSet.eps[i].fqdn, epSet.eps[i].port);
-    if (strcmp(epSet.eps[i].fqdn, tsLocalFqdn) == 0 && epSet.eps[i].port == tsServerPort) {
-      epSet.inUse = (i + 1) % epSet.numOfEps;
+    int32_t contLen = tSerializeSEpSet(NULL, 0, &epSet);
+    pMsg->info.rsp = rpcMallocCont(contLen);
+    if (pMsg->info.rsp != NULL) {
+      tSerializeSEpSet(pMsg->info.rsp, contLen, &epSet);
+      pMsg->info.rspLen = contLen;
+      terrno = TSDB_CODE_RPC_REDIRECT;
+    } else {
+      terrno = TSDB_CODE_OUT_OF_MEMORY;
     }
   }
-#endif  
-
-  int32_t contLen = tSerializeSEpSet(NULL, 0, &epSet);
-  pMsg->info.rsp = rpcMallocCont(contLen);
-  if (pMsg->info.rsp != NULL) {
-    tSerializeSEpSet(pMsg->info.rsp, contLen, &epSet);
-    pMsg->info.rspLen = contLen;
-    terrno = TSDB_CODE_RPC_REDIRECT;
-  } else {
-    terrno = TSDB_CODE_OUT_OF_MEMORY;
-  }
 
   return -1;
 }
 
-static int32_t mndCheckRequestValid(SRpcMsg *pMsg) {
+static int32_t mndCheckMsgContent(SRpcMsg *pMsg) {
   if (!IsReq(pMsg)) return 0;
   if (pMsg->contLen != 0 && pMsg->pCont != NULL) return 0;
 
-  mError("msg:%p, failed to valid request, app:%p type:%s", pMsg, pMsg->info.ahandle, TMSG_INFO(pMsg->msgType));
+  mError("msg:%p, failed to check msg content, app:%p type:%s", pMsg, pMsg->info.ahandle, TMSG_INFO(pMsg->msgType));
   terrno = TSDB_CODE_INVALID_MSG_LEN;
   return -1;
 }
 
-int32_t mndProcessMsg(SRpcMsg *pMsg) {
-  if (mndCheckMnodeMaster(pMsg) != 0) return -1;
-  if (mndCheckRequestValid(pMsg) != 0) return -1;
-
+int32_t mndProcessRpcMsg(SRpcMsg *pMsg) {
   SMnode  *pMnode = pMsg->info.node;
   MndMsgFp fp = pMnode->msgFp[TMSG_INDEX(pMsg->msgType)];
   if (fp == NULL) {
@@ -466,8 +460,13 @@ int32_t mndProcessMsg(SRpcMsg *pMsg) {
     return -1;
   }
 
-  mTrace("msg:%p, will be processed in mnode, app:%p type:%s", pMsg, pMsg->info.ahandle, TMSG_INFO(pMsg->msgType));
+  if (mndCheckMsgContent(pMsg) != 0) return -1;
+  if (mndCheckMnodeState(pMsg) != 0) return -1;
+
+  mTrace("msg:%p, start to process in mnode, app:%p type:%s", pMsg, pMsg->info.ahandle, TMSG_INFO(pMsg->msgType));
   int32_t code = (*fp)(pMsg);
+  mndReleaseRpcRef(pMnode);
+
   if (code == TSDB_CODE_ACTION_IN_PROGRESS) {
     mTrace("msg:%p, won't response immediately since in progress", pMsg);
   } else if (code == 0) {
@@ -476,6 +475,7 @@ int32_t mndProcessMsg(SRpcMsg *pMsg) {
     mError("msg:%p, failed to process since %s, app:%p type:%s", pMsg, terrstr(), pMsg->info.ahandle,
            TMSG_INFO(pMsg->msgType));
   }
+
   return code;
 }
 
@@ -502,7 +502,7 @@ int64_t mndGenerateUid(char *name, int32_t len) {
 
 int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgroupInfo *pVgroupInfo,
                           SMonGrantInfo *pGrantInfo) {
-  if (!mndIsMaster(pMnode)) return -1;
+  if (mndAcquireRpcRef(pMnode) != 0) return -1;
 
   SSdb   *pSdb = pMnode->pSdb;
   int64_t ms = taosGetTimestampMs();
@@ -511,6 +511,7 @@ int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgr
   pClusterInfo->mnodes = taosArrayInit(sdbGetSize(pSdb, SDB_MNODE), sizeof(SMonMnodeDesc));
   pVgroupInfo->vgroups = taosArrayInit(sdbGetSize(pSdb, SDB_VGROUP), sizeof(SMonVgroupDesc));
   if (pClusterInfo->dnodes == NULL || pClusterInfo->mnodes == NULL || pVgroupInfo->vgroups == NULL) {
+    mndReleaseRpcRef(pMnode);
     return -1;
   }
 
@@ -605,6 +606,7 @@ int32_t mndGetMonitorInfo(SMnode *pMnode, SMonClusterInfo *pClusterInfo, SMonVgr
     pGrantInfo->timeseries_total = INT32_MAX;
   }
 
+  mndReleaseRpcRef(pMnode);
   return 0;
 }
 
@@ -612,3 +614,76 @@ int32_t mndGetLoad(SMnode *pMnode, SMnodeLoad *pLoad) {
   pLoad->syncState = syncGetMyRole(pMnode->syncMgmt.sync);
   return 0;
 }
+
+int32_t mndAcquireRpcRef(SMnode *pMnode) {
+  int32_t code = 0;
+  taosThreadRwlockRdlock(&pMnode->lock);
+  if (pMnode->stopped) {
+    terrno = TSDB_CODE_APP_NOT_READY;
+    code = -1;
+  } else if (!mndIsMaster(pMnode)) {
+    code = -1;
+  } else {
+    int32_t ref = atomic_add_fetch_32(&pMnode->rpcRef, 1);
+    mTrace("mnode rpc is acquired, ref:%d", ref);
+  }
+  taosThreadRwlockUnlock(&pMnode->lock);
+  return code;
+}
+
+void mndReleaseRpcRef(SMnode *pMnode) {
+  taosThreadRwlockRdlock(&pMnode->lock);
+  int32_t ref = atomic_sub_fetch_32(&pMnode->rpcRef, 1);
+  mTrace("mnode rpc is released, ref:%d", ref);
+  taosThreadRwlockUnlock(&pMnode->lock);
+}
+
+void mndSetRestore(SMnode *pMnode, bool restored) {
+  if (restored) {
+    taosThreadRwlockWrlock(&pMnode->lock);
+    pMnode->restored = true;
+    taosThreadRwlockUnlock(&pMnode->lock);
+    mTrace("mnode set restored:%d", restored);
+  } else {
+    taosThreadRwlockWrlock(&pMnode->lock);
+    pMnode->restored = false;
+    taosThreadRwlockUnlock(&pMnode->lock);
+    mTrace("mnode set restored:%d", restored);
+    while (1) {
+      if (pMnode->rpcRef <= 0) break;
+      taosMsleep(3);
+    }
+  }
+}
+
+bool mndGetRestored(SMnode *pMnode) { return pMnode->restored; }
+
+void mndSetStop(SMnode *pMnode) {
+  taosThreadRwlockWrlock(&pMnode->lock);
+  pMnode->stopped = true;
+  taosThreadRwlockUnlock(&pMnode->lock);
+  mTrace("mnode set stopped");
+}
+
+bool mndGetStop(SMnode *pMnode) { return pMnode->stopped; }
+
+int32_t mndAcquireSyncRef(SMnode *pMnode) {
+  int32_t code = 0;
+  taosThreadRwlockRdlock(&pMnode->lock);
+  if (pMnode->stopped) {
+    terrno = TSDB_CODE_APP_NOT_READY;
+    code = -1;
+  } else {
+    int32_t ref = atomic_add_fetch_32(&pMnode->syncRef, 1);
+    mTrace("mnode sync is acquired, ref:%d", ref);
+  }
+  taosThreadRwlockUnlock(&pMnode->lock);
+  return code;
+}
+
+void mndReleaseSyncRef(SMnode *pMnode) {
+  taosThreadRwlockRdlock(&pMnode->lock);
+  int32_t ref = atomic_sub_fetch_32(&pMnode->syncRef, 1);
+  mTrace("mnode sync is released, ref:%d", ref);
+  taosThreadRwlockUnlock(&pMnode->lock);
+}
\ No newline at end of file

source/dnode/mnode/impl/src/mndSubscribe.c

@@ -940,7 +940,7 @@ static int32_t mndRetrieveSubscribe(SRpcMsg *pReq, SShowObj *pShow, SSDataBlock
     }
 
     // do not show for cleared subscription
-#if 0
+#if 1
     int32_t sz = taosArrayGetSize(pSub->unassignedVgs);
     for (int32_t i = 0; i < sz; i++) {
       SMqVgEp *pVgEp = taosArrayGetP(pSub->unassignedVgs, i);

source/dnode/mnode/impl/src/mndSync.c

@@ -17,12 +17,12 @@
 #include "mndSync.h"
 #include "mndTrans.h"
 
-int32_t mndSyncEqMsg(const SMsgCb *msgcb, SRpcMsg *pMsg) { 
+int32_t mndSyncEqMsg(const SMsgCb *msgcb, SRpcMsg *pMsg) {
   SMsgHead *pHead = pMsg->pCont;
   pHead->contLen = htonl(pHead->contLen);
   pHead->vgId = htonl(pHead->vgId);
 
-  return tmsgPutToQueue(msgcb, SYNC_QUEUE, pMsg); 
+  return tmsgPutToQueue(msgcb, SYNC_QUEUE, pMsg);
 }
 
 int32_t mndSyncSendMsg(const SEpSet *pEpSet, SRpcMsg *pMsg) { return tmsgSendReq(pEpSet, pMsg); }
@@ -32,7 +32,7 @@ void mndSyncCommitMsg(struct SSyncFSM *pFsm, const SRpcMsg *pMsg, SFsmCbMeta cbM
   SSyncMgmt *pMgmt = &pMnode->syncMgmt;
   SSdbRaw   *pRaw = pMsg->pCont;
 
-  int32_t transId = sdbGetIdFromRaw(pRaw);
+  int32_t transId = sdbGetIdFromRaw(pMnode->pSdb, pRaw);
   pMgmt->errCode = cbMeta.code;
   mTrace("trans:%d, is proposed, savedTransId:%d code:0x%x, ver:%" PRId64 " term:%" PRId64 " role:%s raw:%p", transId,
          pMgmt->transId, cbMeta.code, cbMeta.index, cbMeta.term, syncStr(cbMeta.state), pRaw);
@@ -63,28 +63,41 @@ void mndRestoreFinish(struct SSyncFSM *pFsm) {
   if (!pMnode->deploy) {
     mInfo("mnode sync restore finished");
     mndTransPullup(pMnode);
-    pMnode->syncMgmt.restored = true;
+    mndSetRestore(pMnode, true);
   }
 }
 
-int32_t mndSnapshotRead(struct SSyncFSM* pFsm, const SSnapshot* pSnapshot, void** ppIter, char** ppBuf, int32_t* len) {
-  /*
+int32_t mndSnapshotRead(struct SSyncFSM *pFsm, const SSnapshot *pSnapshot, void **ppIter, char **ppBuf, int32_t *len) {
   SMnode *pMnode = pFsm->data;
-  SSdbIter *pIter;
-  if (iter == NULL) { 
-    pIter = sdbIterInit(pMnode->sdb)
+  mInfo("start to read snapshot from sdb");
+
+  int32_t code = sdbReadSnapshot(pMnode->pSdb, (SSdbIter **)ppIter, ppBuf, len);
+  if (code != 0) {
+    mError("failed to read snapshot from sdb since %s", terrstr());
   } else {
-    pIter = iter;
+    if (*ppIter == NULL) {
+      mInfo("successfully to read snapshot from sdb");
+    }
   }
-  */
 
-  return 0;
+  return code;
 }
 
-int32_t mndSnapshotApply(struct SSyncFSM* pFsm, const SSnapshot* pSnapshot, char* pBuf, int32_t len) {
+int32_t mndSnapshotApply(struct SSyncFSM *pFsm, const SSnapshot *pSnapshot, char *pBuf, int32_t len) {
   SMnode *pMnode = pFsm->data;
-  sdbWrite(pMnode->pSdb, (SSdbRaw*)pBuf);
-  return 0;
+  mndSetRestore(pMnode, false);
+  mInfo("start to apply snapshot to sdb, len:%d", len);
+
+  int32_t code = sdbApplySnapshot(pMnode->pSdb, pBuf, len);
+  if (code != 0) {
+    mError("failed to apply snapshot to sdb, len:%d", len);
+  } else {
+    mInfo("successfully to apply snapshot to sdb, len:%d", len);
+    mndSetRestore(pMnode, true);
+  }
+
+  // taosMemoryFree(pBuf);
+  return code;
 }
 
 void mndReConfig(struct SSyncFSM *pFsm, SSyncCfg newCfg, SReConfigCbMeta cbMeta) {
@@ -116,7 +129,7 @@ SSyncFSM *mndSyncMakeFsm(SMnode *pMnode) {
   pFsm->FpSnapshotRead = mndSnapshotRead;
   pFsm->FpSnapshotApply = mndSnapshotApply;
   pFsm->FpReConfigCb = mndReConfig;
-  
+
   return pFsm;
 }
 
@@ -150,8 +163,7 @@ int32_t mndInitSync(SMnode *pMnode) {
   SSyncCfg *pCfg = &syncInfo.syncCfg;
   pCfg->replicaNum = pMnode->replica;
   pCfg->myIndex = pMnode->selfIndex;
-  mInfo("start to open mnode sync, replica:%d myindex:%d standby:%d", pCfg->replicaNum, pCfg->myIndex,
-        pMgmt->standby);
+  mInfo("start to open mnode sync, replica:%d myindex:%d standby:%d", pCfg->replicaNum, pCfg->myIndex, pMgmt->standby);
   for (int32_t i = 0; i < pMnode->replica; ++i) {
     SNodeInfo *pNode = &pCfg->nodeInfo[i];
     tstrncpy(pNode->nodeFqdn, pMnode->replicas[i].fqdn, sizeof(pNode->nodeFqdn));
@@ -238,7 +250,7 @@ bool mndIsMaster(SMnode *pMnode) {
     return false;
   }
 
-  if (!pMgmt->restored) {
+  if (!pMnode->restored) {
     terrno = TSDB_CODE_APP_NOT_READY;
     return false;
   }

source/dnode/mnode/sdb/inc/sdb.h

@@ -166,7 +166,6 @@ typedef struct SSdbRow {
 typedef struct SSdb {
   SMnode        *pMnode;
   char          *currDir;
-  char          *syncDir;
   char          *tmpDir;
   int64_t        lastCommitVer;
   int64_t        curVer;
@@ -182,11 +181,12 @@ typedef struct SSdb {
   SdbDeployFp    deployFps[SDB_MAX];
   SdbEncodeFp    encodeFps[SDB_MAX];
   SdbDecodeFp    decodeFps[SDB_MAX];
+  TdThreadMutex  filelock;
 } SSdb;
 
 typedef struct SSdbIter {
   TdFilePtr file;
-  int64_t   readlen;
+  int64_t   total;
 } SSdbIter;
 
 typedef struct {
@@ -380,13 +380,12 @@ SSdbRow *sdbAllocRow(int32_t objSize);
 void    *sdbGetRowObj(SSdbRow *pRow);
 void     sdbFreeRow(SSdb *pSdb, SSdbRow *pRow, bool callFunc);
 
-SSdbIter *sdbIterInit(SSdb *pSdb);
-SSdbIter *sdbIterRead(SSdb *pSdb, SSdbIter *iter, char **ppBuf, int32_t *len);
+int32_t sdbReadSnapshot(SSdb *pSdb, SSdbIter **ppIter, char **ppBuf, int32_t *len);
+int32_t sdbApplySnapshot(SSdb *pSdb, char *pBuf, int32_t len);
 
 const char *sdbTableName(ESdbType type);
 void        sdbPrintOper(SSdb *pSdb, SSdbRow *pRow, const char *oper);
-
-int32_t sdbGetIdFromRaw(SSdbRaw *pRaw);
+int32_t     sdbGetIdFromRaw(SSdb *pSdb, SSdbRaw *pRaw);
 
 #ifdef __cplusplus
 }

source/dnode/mnode/sdb/inc/sdbInt.h

-/*
- * Copyright (c) 2019 TAOS Data, Inc. <jhtao@taosdata.com>
- *
- * This program is free software: you can use, redistribute, and/or modify
- * it under the terms of the GNU Affero General Public License, version 3
- * or later ("AGPL"), as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE.
- *
- * You should have received a copy of the GNU Affero General Public License
- * along with this program. If not, see <http://www.gnu.org/licenses/>.
- */
-
-#ifndef _TD_SDB_INT_H_
-#define _TD_SDB_INT_H_
-
-#include "os.h"
-
-#include "sdb.h"
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-// clang-format off
-#define mFatal(...) { if (mDebugFlag & DEBUG_FATAL) { taosPrintLog("MND FATAL ", DEBUG_FATAL, 255, __VA_ARGS__); }}
-#define mError(...) { if (mDebugFlag & DEBUG_ERROR) { taosPrintLog("MND ERROR ", DEBUG_ERROR, 255, __VA_ARGS__); }}
-#define mWarn(...)  { if (mDebugFlag & DEBUG_WARN)  { taosPrintLog("MND WARN ", DEBUG_WARN, 255, __VA_ARGS__); }}
-#define mInfo(...)  { if (mDebugFlag & DEBUG_INFO)  { taosPrintLog("MND ", DEBUG_INFO, 255, __VA_ARGS__); }}
-#define mDebug(...) { if (mDebugFlag & DEBUG_DEBUG) { taosPrintLog("MND ", DEBUG_DEBUG, mDebugFlag, __VA_ARGS__); }}
-#define mTrace(...) { if (mDebugFlag & DEBUG_TRACE) { taosPrintLog("MND ", DEBUG_TRACE, mDebugFlag, __VA_ARGS__); }}
-// clang-format on
-
-typedef struct SSdbRaw {
-  int8_t  type;
-  int8_t  status;
-  int8_t  sver;
-  int8_t  reserved;
-  int32_t dataLen;
-  char    pData[];
-} SSdbRaw;
-
-typedef struct SSdbRow {
-  ESdbType   type;
-  ESdbStatus status;
-  int32_t    refCount;
-  char       pObj[];
-} SSdbRow;
-
-const char *sdbTableName(ESdbType type);
-void        sdbPrintOper(SSdb *pSdb, SSdbRow *pRow, const char *oper);
-
-void sdbFreeRow(SSdb *pSdb, SSdbRow *pRow, bool callFunc);
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif /*_TD_SDB_INT_H_*/

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

@@ -56,6 +56,7 @@ SSdb *sdbInit(SSdbOpt *pOption) {
   pSdb->curTerm = -1;
   pSdb->lastCommitVer = -1;
   pSdb->pMnode = pOption->pMnode;
+  taosThreadMutexInit(&pSdb->filelock, NULL);
   mDebug("sdb init successfully");
   return pSdb;
 }
@@ -69,10 +70,6 @@ void sdbCleanup(SSdb *pSdb) {
     taosMemoryFreeClear(pSdb->currDir);
   }
 
-  if (pSdb->syncDir != NULL) {
-    taosMemoryFreeClear(pSdb->syncDir);
-  }
-
   if (pSdb->tmpDir != NULL) {
     taosMemoryFreeClear(pSdb->tmpDir);
   }
@@ -104,6 +101,7 @@ void sdbCleanup(SSdb *pSdb) {
     mDebug("sdb table:%s is cleaned up", sdbTableName(i));
   }
 
+  taosThreadMutexDestroy(&pSdb->filelock);
   taosMemoryFree(pSdb);
   mDebug("sdb is cleaned up");
 }

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

@@ -22,13 +22,14 @@
 #define SDB_RESERVE_SIZE 512
 #define SDB_FILE_VER     1
 
-static int32_t sdbRunDeployFp(SSdb *pSdb) {
+static int32_t sdbDeployData(SSdb *pSdb) {
   mDebug("start to deploy sdb");
 
   for (int32_t i = SDB_MAX - 1; i >= 0; --i) {
     SdbDeployFp fp = pSdb->deployFps[i];
     if (fp == NULL) continue;
 
+    mDebug("start to deploy sdb:%s", sdbTableName(i));
     if ((*fp)(pSdb->pMnode) != 0) {
       mError("failed to deploy sdb:%s since %s", sdbTableName(i), terrstr());
       return -1;
@@ -39,6 +40,39 @@ static int32_t sdbRunDeployFp(SSdb *pSdb) {
   return 0;
 }
 
+static void sdbResetData(SSdb *pSdb) {
+  mDebug("start to reset sdb");
+
+  for (ESdbType i = 0; i < SDB_MAX; ++i) {
+    SHashObj *hash = pSdb->hashObjs[i];
+    if (hash == NULL) continue;
+
+    SSdbRow **ppRow = taosHashIterate(hash, NULL);
+    while (ppRow != NULL) {
+      SSdbRow *pRow = *ppRow;
+      if (pRow == NULL) continue;
+
+      sdbFreeRow(pSdb, pRow, true);
+      ppRow = taosHashIterate(hash, ppRow);
+    }
+  }
+
+  for (ESdbType i = 0; i < SDB_MAX; ++i) {
+    SHashObj *hash = pSdb->hashObjs[i];
+    if (hash == NULL) continue;
+
+    taosHashClear(pSdb->hashObjs[i]);
+    pSdb->tableVer[i] = 0;
+    pSdb->maxId[i] = 0;
+    mDebug("sdb:%s is reset", sdbTableName(i));
+  }
+
+  pSdb->curVer = -1;
+  pSdb->curTerm = -1;
+  pSdb->lastCommitVer = -1;
+  mDebug("sdb reset successfully");
+}
+
 static int32_t sdbReadFileHead(SSdb *pSdb, TdFilePtr pFile) {
   int64_t sver = 0;
   int32_t ret = taosReadFile(pFile, &sver, sizeof(int64_t));
@@ -169,11 +203,15 @@ static int32_t sdbWriteFileHead(SSdb *pSdb, TdFilePtr pFile) {
   return 0;
 }
 
-int32_t sdbReadFile(SSdb *pSdb) {
+static int32_t sdbReadFileImp(SSdb *pSdb) {
   int64_t offset = 0;
   int32_t code = 0;
   int32_t readLen = 0;
   int64_t ret = 0;
+  char    file[PATH_MAX] = {0};
+
+  snprintf(file, sizeof(file), "%s%ssdb.data", pSdb->currDir, TD_DIRSEP);
+  mDebug("start to read file:%s", file);
 
   SSdbRaw *pRaw = taosMemoryMalloc(WAL_MAX_SIZE + 100);
   if (pRaw == NULL) {
@@ -182,10 +220,6 @@ int32_t sdbReadFile(SSdb *pSdb) {
     return -1;
   }
 
-  char file[PATH_MAX] = {0};
-  snprintf(file, sizeof(file), "%s%ssdb.data", pSdb->currDir, TD_DIRSEP);
-  mDebug("start to read file:%s", file);
-
   TdFilePtr pFile = taosOpenFile(file, TD_FILE_READ);
   if (pFile == NULL) {
     taosMemoryFree(pRaw);
@@ -196,8 +230,6 @@ int32_t sdbReadFile(SSdb *pSdb) {
 
   if (sdbReadFileHead(pSdb, pFile) != 0) {
     mError("failed to read file:%s head since %s", file, terrstr());
-    pSdb->curVer = -1;
-    pSdb->curTerm = -1;
     taosMemoryFree(pRaw);
     taosCloseFile(&pFile);
     return -1;
@@ -264,6 +296,20 @@ _OVER:
   return code;
 }
 
+int32_t sdbReadFile(SSdb *pSdb) {
+  taosThreadMutexLock(&pSdb->filelock);
+
+  sdbResetData(pSdb);
+  int32_t code = sdbReadFileImp(pSdb);
+  if (code != 0) {
+    mError("failed to read sdb since %s", terrstr());
+    sdbResetData(pSdb);
+  }
+
+  taosThreadMutexUnlock(&pSdb->filelock);
+  return code;
+}
+
 static int32_t sdbWriteFileImp(SSdb *pSdb) {
   int32_t code = 0;
 
@@ -378,32 +424,41 @@ int32_t sdbWriteFile(SSdb *pSdb) {
     return 0;
   }
 
-  return sdbWriteFileImp(pSdb);
+  taosThreadMutexLock(&pSdb->filelock);
+  int32_t code = sdbWriteFileImp(pSdb);
+  if (code != 0) {
+    mError("failed to write sdb since %s", terrstr());
+  }
+  taosThreadMutexUnlock(&pSdb->filelock);
+  return code;
 }
 
 int32_t sdbDeploy(SSdb *pSdb) {
-  if (sdbRunDeployFp(pSdb) != 0) {
+  if (sdbDeployData(pSdb) != 0) {
     return -1;
   }
 
-  if (sdbWriteFileImp(pSdb) != 0) {
+  if (sdbWriteFile(pSdb) != 0) {
     return -1;
   }
 
   return 0;
 }
 
-SSdbIter *sdbIterInit(SSdb *pSdb) {
+static SSdbIter *sdbOpenIter(SSdb *pSdb) {
   char datafile[PATH_MAX] = {0};
   char tmpfile[PATH_MAX] = {0};
   snprintf(datafile, sizeof(datafile), "%s%ssdb.data", pSdb->currDir, TD_DIRSEP);
-  snprintf(tmpfile, sizeof(datafile), "%s%ssdb.data", pSdb->tmpDir, TD_DIRSEP);
+  snprintf(tmpfile, sizeof(tmpfile), "%s%ssdb.data", pSdb->tmpDir, TD_DIRSEP);
 
+  taosThreadMutexLock(&pSdb->filelock);
   if (taosCopyFile(datafile, tmpfile) != 0) {
+    taosThreadMutexUnlock(&pSdb->filelock);
     terrno = TAOS_SYSTEM_ERROR(errno);
     mError("failed to copy file %s to %s since %s", datafile, tmpfile, terrstr());
     return NULL;
   }
+  taosThreadMutexUnlock(&pSdb->filelock);
 
   SSdbIter *pIter = taosMemoryCalloc(1, sizeof(SSdbIter));
   if (pIter == NULL) {
@@ -414,44 +469,144 @@ SSdbIter *sdbIterInit(SSdb *pSdb) {
   pIter->file = taosOpenFile(tmpfile, TD_FILE_READ);
   if (pIter->file == NULL) {
     terrno = TAOS_SYSTEM_ERROR(errno);
-    mError("failed to read snapshot file:%s since %s", tmpfile, terrstr());
+    mError("failed to read file:%s since %s", tmpfile, terrstr());
     taosMemoryFree(pIter);
     return NULL;
   }
 
-  mDebug("start to read snapshot file:%s, iter:%p", tmpfile, pIter);
   return pIter;
 }
 
-SSdbIter *sdbIterRead(SSdb *pSdb, SSdbIter *pIter, char **ppBuf, int32_t *buflen) {
-  const int32_t maxlen = 100;
+static void sdbCloseIter(SSdb *pSdb, SSdbIter *pIter) {
+  if (pIter == NULL) return;
+  if (pIter->file != NULL) {
+    taosCloseFile(&pIter->file);
+  }
+
+  char tmpfile[PATH_MAX] = {0};
+  snprintf(tmpfile, sizeof(tmpfile), "%s%ssdb.data", pSdb->tmpDir, TD_DIRSEP);
+  taosRemoveFile(tmpfile);
+
+  taosMemoryFree(pIter);
+  mInfo("sdbiter:%p, is closed", pIter);
+}
+
+static SSdbIter *sdbGetIter(SSdb *pSdb, SSdbIter **ppIter) {
+  SSdbIter *pIter = NULL;
+  if (ppIter != NULL) pIter = *ppIter;
+
+  if (pIter == NULL) {
+    pIter = sdbOpenIter(pSdb);
+    if (pIter != NULL) {
+      mInfo("sdbiter:%p, is created to read snapshot", pIter);
+      *ppIter = pIter;
+    } else {
+      mError("failed to create sdbiter to read snapshot since %s", terrstr());
+      *ppIter = NULL;
+      return NULL;
+    }
+  } else {
+    mInfo("sdbiter:%p, continue to read snapshot, total:%" PRId64, pIter, pIter->total);
+  }
+
+  return pIter;
+}
+
+int32_t sdbReadSnapshot(SSdb *pSdb, SSdbIter **ppIter, char **ppBuf, int32_t *len) {
+  SSdbIter *pIter = sdbGetIter(pSdb, ppIter);
+  if (pIter == NULL) return -1;
 
-  char *pBuf = taosMemoryCalloc(1, maxlen);
+  int32_t maxlen = 100;
+  char   *pBuf = taosMemoryCalloc(1, maxlen);
   if (pBuf == NULL) {
     terrno = TSDB_CODE_OUT_OF_MEMORY;
-    return NULL;
+    sdbCloseIter(pSdb, pIter);
+    return -1;
   }
 
   int32_t readlen = taosReadFile(pIter->file, pBuf, maxlen);
-  if (readlen == 0) {
-    mTrace("read snapshot to the end, readlen:%" PRId64, pIter->readlen);
-    taosMemoryFree(pBuf);
-    taosCloseFile(&pIter->file);
-    taosMemoryFree(pIter);
-    pIter = NULL;
-  } else if (readlen < 0) {
+  if (readlen < 0 || (readlen == 0 && errno != 0)) {
     terrno = TAOS_SYSTEM_ERROR(errno);
-    mError("failed to read snapshot since %s, readlen:%" PRId64, terrstr(), pIter->readlen);
+    mError("sdbiter:%p, failed to read snapshot since %s, total:%" PRId64, pIter, terrstr(), pIter->total);
+    *ppBuf = NULL;
+    *len = 0;
+    *ppIter = NULL;
+    sdbCloseIter(pSdb, pIter);
     taosMemoryFree(pBuf);
-    taosCloseFile(&pIter->file);
-    taosMemoryFree(pIter);
-    pIter = NULL;
-  } else {
-    pIter->readlen += readlen;
-    mTrace("read snapshot, readlen:%" PRId64, pIter->readlen);
+    return -1;
+  } else if (readlen == 0) {
+    mInfo("sdbiter:%p, read snapshot to the end, total:%" PRId64, pIter, pIter->total);
+    *ppBuf = NULL;
+    *len = 0;
+    *ppIter = NULL;
+    sdbCloseIter(pSdb, pIter);
+    taosMemoryFree(pBuf);
+    return 0;
+  } else if ((readlen < maxlen && errno != 0) || readlen == maxlen) {
+    pIter->total += readlen;
+    mInfo("sdbiter:%p, read:%d bytes from snapshot, total:%" PRId64, pIter, readlen, pIter->total);
+    *ppBuf = pBuf;
+    *len = readlen;
+    return 0;
+  } else if (readlen < maxlen && errno == 0) {
+    mInfo("sdbiter:%p, read snapshot to the end, total:%" PRId64, pIter, pIter->total);
     *ppBuf = pBuf;
-    *buflen = readlen;
+    *len = readlen;
+    *ppIter = NULL;
+    sdbCloseIter(pSdb, pIter);
+    return 0;
+  } else {
+    // impossible
+    mError("sdbiter:%p, read:%d bytes from snapshot, total:%" PRId64, pIter, readlen, pIter->total);
+    *ppBuf = NULL;
+    *len = 0;
+    *ppIter = NULL;
+    sdbCloseIter(pSdb, pIter);
+    taosMemoryFree(pBuf);
+    return -1;
   }
-
-  return pIter;
 }
+
+int32_t sdbApplySnapshot(SSdb *pSdb, char *pBuf, int32_t len) {
+  char datafile[PATH_MAX] = {0};
+  char tmpfile[PATH_MAX] = {0};
+  snprintf(datafile, sizeof(datafile), "%s%ssdb.data", pSdb->currDir, TD_DIRSEP);
+  snprintf(tmpfile, sizeof(tmpfile), "%s%ssdb.data", pSdb->tmpDir, TD_DIRSEP);
+
+  TdFilePtr pFile = taosOpenFile(tmpfile, TD_FILE_CREATE | TD_FILE_WRITE | TD_FILE_TRUNC);
+  if (pFile == NULL) {
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    mError("failed to write %s since %s", tmpfile, terrstr());
+    return -1;
+  }
+
+  int32_t writelen = taosWriteFile(pFile, pBuf, len);
+  if (writelen != len) {
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    mError("failed to write %s since %s", tmpfile, terrstr());
+    taosCloseFile(&pFile);
+    return -1;
+  }
+
+  if (taosFsyncFile(pFile) != 0) {
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    mError("failed to fsync %s since %s", tmpfile, terrstr());
+    taosCloseFile(&pFile);
+    return -1;
+  }
+
+  (void)taosCloseFile(&pFile);
+
+  if (taosRenameFile(tmpfile, datafile) != 0) {
+    terrno = TAOS_SYSTEM_ERROR(errno);
+    mError("failed to rename file %s to %s since %s", tmpfile, datafile, terrstr());
+    return -1;
+  }
+
+  if (sdbReadFile(pSdb) != 0) {
+    mError("failed to read from %s since %s", datafile, terrstr());
+    return -1;
+  }
+
+  return 0;
+}
\ No newline at end of file

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

@@ -16,9 +16,14 @@
 #define _DEFAULT_SOURCE
 #include "sdb.h"
 
-int32_t sdbGetIdFromRaw(SSdbRaw *pRaw) {
-  int32_t id = *((int32_t *)(pRaw->pData));
-  return id;
+int32_t sdbGetIdFromRaw(SSdb *pSdb, SSdbRaw *pRaw) {
+  EKeyType keytype = pSdb->keyTypes[pRaw->type];
+  if (keytype == SDB_KEY_INT32) {
+    int32_t id = *((int32_t *)(pRaw->pData));
+    return id;
+  } else {
+    return -2;
+  }
 }
 
 SSdbRaw *sdbAllocRaw(ESdbType type, int8_t sver, int32_t dataLen) {

source/dnode/vnode/inc/vnode.h

@@ -105,13 +105,15 @@ tsdbReaderT  tsdbQueryCacheLast(SVnode *pVnode, SQueryTableDataCond *pCond, STab
                                 void *pMemRef);
 int32_t      tsdbGetFileBlocksDistInfo(tsdbReaderT *pReader, STableBlockDistInfo *pTableBlockInfo);
 bool         isTsdbCacheLastRow(tsdbReaderT *pReader);
-int32_t      tsdbGetAllTableList(SMeta* pMeta, uint64_t uid, SArray* list);
+int32_t      tsdbGetAllTableList(SMeta *pMeta, uint64_t uid, SArray *list);
+void *       tsdbGetIdx(SMeta *pMeta);
 int64_t      tsdbGetNumOfRowsInMemTable(tsdbReaderT *pHandle);
-bool         tsdbNextDataBlock(tsdbReaderT pTsdbReadHandle);
-void         tsdbRetrieveDataBlockInfo(tsdbReaderT *pTsdbReadHandle, SDataBlockInfo *pBlockInfo);
+
+bool    tsdbNextDataBlock(tsdbReaderT pTsdbReadHandle);
+void    tsdbRetrieveDataBlockInfo(tsdbReaderT *pTsdbReadHandle, SDataBlockInfo *pBlockInfo);
 int32_t tsdbRetrieveDataBlockStatisInfo(tsdbReaderT *pTsdbReadHandle, SColumnDataAgg ***pBlockStatis, bool *allHave);
 SArray *tsdbRetrieveDataBlock(tsdbReaderT *pTsdbReadHandle, SArray *pColumnIdList);
-void    tsdbResetReadHandle(tsdbReaderT queryHandle, SQueryTableDataCond *pCond);
+void    tsdbResetReadHandle(tsdbReaderT queryHandle, SQueryTableDataCond *pCond, int32_t tWinIdx);
 void    tsdbCleanupReadHandle(tsdbReaderT queryHandle);
 
 // tq
@@ -174,7 +176,7 @@ struct SMetaEntry {
   int64_t  version;
   int8_t   type;
   tb_uid_t uid;
-  char    *name;
+  char *   name;
   union {
     struct {
       SSchemaWrapper schemaRow;
@@ -202,17 +204,17 @@ struct SMetaEntry {
 
 struct SMetaReader {
   int32_t    flags;
-  SMeta     *pMeta;
+  SMeta *    pMeta;
   SDecoder   coder;
   SMetaEntry me;
-  void      *pBuf;
+  void *     pBuf;
   int32_t    szBuf;
 };
 
 struct SMTbCursor {
-  TBC        *pDbc;
-  void       *pKey;
-  void       *pVal;
+  TBC *       pDbc;
+  void *      pKey;
+  void *      pVal;
   int32_t     kLen;
   int32_t     vLen;
   SMetaReader mr;

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

@@ -103,6 +103,7 @@ SArray*         metaGetSmaTbUids(SMeta* pMeta);
 int32_t         metaSnapshotReaderOpen(SMeta* pMeta, SMetaSnapshotReader** ppReader, int64_t sver, int64_t ever);
 int32_t         metaSnapshotReaderClose(SMetaSnapshotReader* pReader);
 int32_t         metaSnapshotRead(SMetaSnapshotReader* pReader, void** ppData, uint32_t* nData);
+void*           metaGetIdx(SMeta* pMeta);
 
 int32_t metaCreateTSma(SMeta* pMeta, int64_t version, SSmaCfg* pCfg);
 int32_t metaDropTSma(SMeta* pMeta, int64_t indexUid);

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

@@ -155,44 +155,52 @@ int metaTbCursorNext(SMTbCursor *pTbCur) {
 }
 
 SSchemaWrapper *metaGetTableSchema(SMeta *pMeta, tb_uid_t uid, int32_t sver, bool isinline) {
-  void           *pKey = NULL;
-  void           *pVal = NULL;
-  int             kLen = 0;
-  int             vLen = 0;
-  int             ret;
-  SSkmDbKey       skmDbKey;
-  SSchemaWrapper *pSW = NULL;
-  SSchema        *pSchema = NULL;
-  void           *pBuf;
-  SDecoder        coder = {0};
-
-  // fetch
-  skmDbKey.uid = uid;
-  skmDbKey.sver = sver;
-  pKey = &skmDbKey;
-  kLen = sizeof(skmDbKey);
+  void           *pData = NULL;
+  int             nData = 0;
+  int64_t         version;
+  SSchemaWrapper  schema = {0};
+  SSchemaWrapper *pSchema = NULL;
+  SDecoder        dc = {0};
+
   metaRLock(pMeta);
-  ret = tdbTbGet(pMeta->pSkmDb, pKey, kLen, &pVal, &vLen);
-  metaULock(pMeta);
-  if (ret < 0) {
-    return NULL;
-  }
+  if (sver < 0) {
+    if (tdbTbGet(pMeta->pUidIdx, &uid, sizeof(uid), &pData, &nData) < 0) {
+      goto _err;
+    }
 
-  // decode
-  pBuf = pVal;
-  pSW = taosMemoryMalloc(sizeof(SSchemaWrapper));
+    version = *(int64_t *)pData;
 
-  tDecoderInit(&coder, pVal, vLen);
-  tDecodeSSchemaWrapper(&coder, pSW);
-  pSchema = taosMemoryMalloc(sizeof(SSchema) * pSW->nCols);
-  memcpy(pSchema, pSW->pSchema, sizeof(SSchema) * pSW->nCols);
-  tDecoderClear(&coder);
+    tdbTbGet(pMeta->pTbDb, &(STbDbKey){.uid = uid, .version = version}, sizeof(STbDbKey), &pData, &nData);
+
+    SMetaEntry me = {0};
+    tDecoderInit(&dc, pData, nData);
+    metaDecodeEntry(&dc, &me);
+    if (me.type == TSDB_SUPER_TABLE) {
+      pSchema = tCloneSSchemaWrapper(&me.stbEntry.schemaRow);
+    } else if (me.type == TSDB_NORMAL_TABLE) {
+    } else {
+      ASSERT(0);
+    }
+    tDecoderClear(&dc);
+  } else {
+    if (tdbTbGet(pMeta->pSkmDb, &(SSkmDbKey){.uid = uid, .sver = sver}, sizeof(SSkmDbKey), &pData, &nData) < 0) {
+      goto _err;
+    }
 
-  pSW->pSchema = pSchema;
+    tDecoderInit(&dc, pData, nData);
+    tDecodeSSchemaWrapper(&dc, &schema);
+    pSchema = tCloneSSchemaWrapper(&schema);
+    tDecoderClear(&dc);
+  }
 
-  tdbFree(pVal);
+  metaULock(pMeta);
+  tdbFree(pData);
+  return pSchema;
 
-  return pSW;
+_err:
+  metaULock(pMeta);
+  tdbFree(pData);
+  return NULL;
 }
 
 struct SMCtbCursor {

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

@@ -31,9 +31,9 @@ int metaCreateSTable(SMeta *pMeta, int64_t version, SVCreateStbReq *pReq) {
   int         vLen = 0;
   const void *pKey = NULL;
   const void *pVal = NULL;
-  void       *pBuf = NULL;
+  void *      pBuf = NULL;
   int32_t     szBuf = 0;
-  void       *p = NULL;
+  void *      p = NULL;
   SMetaReader mr = {0};
 
   // validate req
@@ -87,7 +87,7 @@ int metaDropSTable(SMeta *pMeta, int64_t verison, SVDropStbReq *pReq) {
   }
 
   // drop all child tables
-  TBC    *pCtbIdxc = NULL;
+  TBC *   pCtbIdxc = NULL;
   SArray *pArray = taosArrayInit(8, sizeof(tb_uid_t));
 
   tdbTbcOpen(pMeta->pCtbIdx, &pCtbIdxc, &pMeta->txn);
@@ -142,8 +142,8 @@ _exit:
 int metaAlterSTable(SMeta *pMeta, int64_t version, SVCreateStbReq *pReq) {
   SMetaEntry  oStbEntry = {0};
   SMetaEntry  nStbEntry = {0};
-  TBC        *pUidIdxc = NULL;
-  TBC        *pTbDbc = NULL;
+  TBC *       pUidIdxc = NULL;
+  TBC *       pTbDbc = NULL;
   const void *pData;
   int         nData;
   int64_t     oversion;
@@ -262,7 +262,7 @@ _err:
 }
 
 int metaDropTable(SMeta *pMeta, int64_t version, SVDropTbReq *pReq, SArray *tbUids) {
-  void    *pData = NULL;
+  void *   pData = NULL;
   int      nData = 0;
   int      rc = 0;
   tb_uid_t uid;
@@ -288,7 +288,7 @@ int metaDropTable(SMeta *pMeta, int64_t version, SVDropTbReq *pReq, SArray *tbUi
 }
 
 static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type) {
-  void      *pData = NULL;
+  void *     pData = NULL;
   int        nData = 0;
   int        rc = 0;
   int64_t    version;
@@ -324,14 +324,14 @@ static int metaDropTableByUid(SMeta *pMeta, tb_uid_t uid, int *type) {
 }
 
 static int metaAlterTableColumn(SMeta *pMeta, int64_t version, SVAlterTbReq *pAlterTbReq) {
-  void           *pVal = NULL;
+  void *          pVal = NULL;
   int             nVal = 0;
-  const void     *pData = NULL;
+  const void *    pData = NULL;
   int             nData = 0;
   int             ret = 0;
   tb_uid_t        uid;
   int64_t         oversion;
-  SSchema        *pColumn = NULL;
+  SSchema *       pColumn = NULL;
   SMetaEntry      entry = {0};
   SSchemaWrapper *pSchema;
   int             c;
@@ -479,7 +479,7 @@ _err:
 static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pAlterTbReq) {
   SMetaEntry  ctbEntry = {0};
   SMetaEntry  stbEntry = {0};
-  void       *pVal = NULL;
+  void *      pVal = NULL;
   int         nVal = 0;
   int         ret;
   int         c;
@@ -510,7 +510,7 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
   oversion = *(int64_t *)pData;
 
   // search table.db
-  TBC     *pTbDbc = NULL;
+  TBC *    pTbDbc = NULL;
   SDecoder dc1 = {0};
   SDecoder dc2 = {0};
 
@@ -534,7 +534,7 @@ static int metaUpdateTableTagVal(SMeta *pMeta, int64_t version, SVAlterTbReq *pA
   metaDecodeEntry(&dc2, &stbEntry);
 
   SSchemaWrapper *pTagSchema = &stbEntry.stbEntry.schemaTag;
-  SSchema        *pColumn = NULL;
+  SSchema *       pColumn = NULL;
   int32_t         iCol = 0;
   for (;;) {
     pColumn = NULL;
@@ -639,8 +639,8 @@ int metaAlterTable(SMeta *pMeta, int64_t version, SVAlterTbReq *pReq) {
 
 static int metaSaveToTbDb(SMeta *pMeta, const SMetaEntry *pME) {
   STbDbKey tbDbKey;
-  void    *pKey = NULL;
-  void    *pVal = NULL;
+  void *   pKey = NULL;
+  void *   pVal = NULL;
   int      kLen = 0;
   int      vLen = 0;
   SEncoder coder = {0};
@@ -755,14 +755,14 @@ static void metaDestroyTagIdxKey(STagIdxKey *pTagIdxKey) {
 }
 
 static int metaUpdateTagIdx(SMeta *pMeta, const SMetaEntry *pCtbEntry) {
-  void          *pData = NULL;
+  void *         pData = NULL;
   int            nData = 0;
   STbDbKey       tbDbKey = {0};
   SMetaEntry     stbEntry = {0};
-  STagIdxKey    *pTagIdxKey = NULL;
+  STagIdxKey *   pTagIdxKey = NULL;
   int32_t        nTagIdxKey;
   const SSchema *pTagColumn;       // = &stbEntry.stbEntry.schema.pSchema[0];
-  const void    *pTagData = NULL;  //
+  const void *   pTagData = NULL;  //
   SDecoder       dc = {0};
 
   // get super table
@@ -804,7 +804,7 @@ static int metaUpdateTagIdx(SMeta *pMeta, const SMetaEntry *pCtbEntry) {
 
 static int metaSaveToSkmDb(SMeta *pMeta, const SMetaEntry *pME) {
   SEncoder              coder = {0};
-  void                 *pVal = NULL;
+  void *                pVal = NULL;
   int                   vLen = 0;
   int                   rcode = 0;
   SSkmDbKey             skmDbKey = {0};
@@ -880,3 +880,11 @@ _err:
   metaULock(pMeta);
   return -1;
 }
+// refactor later
+void *metaGetIdx(SMeta *pMeta) {
+#ifdef USE_INVERTED_INDEX
+  return pMeta->pTagIvtIdx;
+#else
+  return pMeta->pTagIdx;
+#endif
+}

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

@@ -235,6 +235,15 @@ int32_t tqUpdateTbUidList(STQ* pTq, const SArray* tbUidList, bool isAdd) {
       }
     }
   }
+  while (1) {
+    pIter = taosHashIterate(pTq->pStreamTasks, pIter);
+    if (pIter == NULL) break;
+    SStreamTask* pTask = (SStreamTask*)pIter;
+    if (pTask->inputType == STREAM_INPUT__DATA_SUBMIT) {
+      int32_t code = qUpdateQualifiedTableId(pTask->exec.executor, tbUidList, isAdd);
+      ASSERT(code == 0);
+    }
+  }
   return 0;
 }
 

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

@@ -678,6 +678,7 @@ static int vnodeProcessSubmitReq(SVnode *pVnode, int64_t version, void *pReq, in
   int32_t        nRows;
   int32_t        tsize, ret;
   SEncoder       encoder = {0};
+  SArray        *newTbUids = NULL;
   terrno = TSDB_CODE_SUCCESS;
 
   pRsp->code = 0;
@@ -698,6 +699,7 @@ static int vnodeProcessSubmitReq(SVnode *pVnode, int64_t version, void *pReq, in
   }
 
   submitRsp.pArray = taosArrayInit(pSubmitReq->numOfBlocks, sizeof(SSubmitBlkRsp));
+  newTbUids = taosArrayInit(pSubmitReq->numOfBlocks, sizeof(int64_t));
   if (!submitRsp.pArray) {
     pRsp->code = TSDB_CODE_OUT_OF_MEMORY;
     goto _exit;
@@ -727,6 +729,7 @@ static int vnodeProcessSubmitReq(SVnode *pVnode, int64_t version, void *pReq, in
           goto _exit;
         }
       }
+      taosArrayPush(newTbUids, &createTbReq.uid);
 
       submitBlkRsp.uid = createTbReq.uid;
       submitBlkRsp.tblFName = taosMemoryMalloc(strlen(pVnode->config.dbname) + strlen(createTbReq.name) + 2);
@@ -754,8 +757,10 @@ static int vnodeProcessSubmitReq(SVnode *pVnode, int64_t version, void *pReq, in
     submitRsp.affectedRows += submitBlkRsp.affectedRows;
     taosArrayPush(submitRsp.pArray, &submitBlkRsp);
   }
+  tqUpdateTbUidList(pVnode->pTq, newTbUids, true);
 
 _exit:
+  taosArrayDestroy(newTbUids);
   tEncodeSize(tEncodeSSubmitRsp, &submitRsp, tsize, ret);
   pRsp->pCont = rpcMallocCont(tsize);
   pRsp->contLen = tsize;

source/libs/catalog/inc/catalogInt.h

@@ -49,19 +49,21 @@ enum {
 };
 
 enum {
-  CTG_ACT_UPDATE_VG = 0,
-  CTG_ACT_UPDATE_TBL,
-  CTG_ACT_REMOVE_DB,
-  CTG_ACT_REMOVE_STB,
-  CTG_ACT_REMOVE_TBL,
-  CTG_ACT_UPDATE_USER,
-  CTG_ACT_MAX
+  CTG_OP_UPDATE_VGROUP = 0,
+  CTG_OP_UPDATE_TB_META,
+  CTG_OP_DROP_DB_CACHE,
+  CTG_OP_DROP_STB_META,
+  CTG_OP_DROP_TB_META,
+  CTG_OP_UPDATE_USER,
+  CTG_OP_UPDATE_VG_EPSET,
+  CTG_OP_MAX
 };
 
 typedef enum {
   CTG_TASK_GET_QNODE = 0,
   CTG_TASK_GET_DB_VGROUP,
   CTG_TASK_GET_DB_CFG,
+  CTG_TASK_GET_DB_INFO,
   CTG_TASK_GET_TB_META,
   CTG_TASK_GET_TB_HASH,
   CTG_TASK_GET_INDEX,
@@ -98,6 +100,10 @@ typedef struct SCtgDbCfgCtx {
   char dbFName[TSDB_DB_FNAME_LEN];
 } SCtgDbCfgCtx;
 
+typedef struct SCtgDbInfoCtx {
+  char dbFName[TSDB_DB_FNAME_LEN];
+} SCtgDbInfoCtx;
+
 typedef struct SCtgTbHashCtx {
   char dbFName[TSDB_DB_FNAME_LEN];
   SName* pName;
@@ -182,6 +188,7 @@ typedef struct SCtgJob {
   int32_t          dbCfgNum;
   int32_t          indexNum;
   int32_t          userNum;
+  int32_t          dbInfoNum;
 } SCtgJob;
 
 typedef struct SCtgMsgCtx {
@@ -285,16 +292,22 @@ typedef struct SCtgUpdateUserMsg {
   SGetUserAuthRsp userAuth;
 } SCtgUpdateUserMsg;
 
+typedef struct SCtgUpdateEpsetMsg {
+  SCatalog* pCtg;
+  char  dbFName[TSDB_DB_FNAME_LEN];
+  int32_t vgId;
+  SEpSet  epSet;
+} SCtgUpdateEpsetMsg;
 
-typedef struct SCtgMetaAction {
-  int32_t  act;
+typedef struct SCtgCacheOperation {
+  int32_t  opId;
   void    *data;
   bool     syncReq;
   uint64_t seqId;
-} SCtgMetaAction;
+} SCtgCacheOperation;
 
 typedef struct SCtgQNode {
-  SCtgMetaAction         action;
+  SCtgCacheOperation     op;
   struct SCtgQNode      *next;
 } SCtgQNode;
 
@@ -321,13 +334,13 @@ typedef struct SCatalogMgmt {
 } SCatalogMgmt;
 
 typedef uint32_t (*tableNameHashFp)(const char *, uint32_t);
-typedef int32_t (*ctgActFunc)(SCtgMetaAction *);
+typedef int32_t (*ctgOpFunc)(SCtgCacheOperation *);
 
-typedef struct SCtgAction {
-  int32_t    actId;
+typedef struct SCtgOperation {
+  int32_t    opId;
   char       name[32];
-  ctgActFunc func;
-} SCtgAction;
+  ctgOpFunc func;
+} SCtgOperation;
 
 #define CTG_QUEUE_ADD() atomic_add_fetch_64(&gCtgMgmt.queue.qRemainNum, 1)
 #define CTG_QUEUE_SUB() atomic_sub_fetch_64(&gCtgMgmt.queue.qRemainNum, 1)
@@ -435,12 +448,13 @@ int32_t ctgdShowCacheInfo(void);
 int32_t ctgRemoveTbMetaFromCache(SCatalog* pCtg, SName* pTableName, bool syncReq);
 int32_t ctgGetTbMetaFromCache(CTG_PARAMS, SCtgTbMetaCtx* ctx, STableMeta** pTableMeta);
 
-int32_t ctgActUpdateVg(SCtgMetaAction *action);
-int32_t ctgActUpdateTb(SCtgMetaAction *action);
-int32_t ctgActRemoveDB(SCtgMetaAction *action);
-int32_t ctgActRemoveStb(SCtgMetaAction *action);
-int32_t ctgActRemoveTb(SCtgMetaAction *action);
-int32_t ctgActUpdateUser(SCtgMetaAction *action);
+int32_t ctgOpUpdateVgroup(SCtgCacheOperation *action);
+int32_t ctgOpUpdateTbMeta(SCtgCacheOperation *action);
+int32_t ctgOpDropDbCache(SCtgCacheOperation *action);
+int32_t ctgOpDropStbMeta(SCtgCacheOperation *action);
+int32_t ctgOpDropTbMeta(SCtgCacheOperation *action);
+int32_t ctgOpUpdateUser(SCtgCacheOperation *action);
+int32_t ctgOpUpdateEpset(SCtgCacheOperation *operation);
 int32_t ctgAcquireVgInfoFromCache(SCatalog* pCtg, const char *dbFName, SCtgDBCache **pCache);
 void ctgReleaseDBCache(SCatalog *pCtg, SCtgDBCache *dbCache);
 void ctgReleaseVgInfo(SCtgDBCache *dbCache);
@@ -449,12 +463,13 @@ int32_t ctgTbMetaExistInCache(SCatalog* pCtg, char *dbFName, char* tbName, int32
 int32_t ctgReadTbMetaFromCache(SCatalog* pCtg, SCtgTbMetaCtx* ctx, STableMeta** pTableMeta);
 int32_t ctgReadTbVerFromCache(SCatalog *pCtg, const SName *pTableName, int32_t *sver, int32_t *tver, int32_t *tbType, uint64_t *suid, char *stbName);
 int32_t ctgChkAuthFromCache(SCatalog* pCtg, const char* user, const char* dbFName, AUTH_TYPE type, bool *inCache, bool *pass);
-int32_t ctgPutRmDBToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId);
-int32_t ctgPutRmStbToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, const char *stbName, uint64_t suid, bool syncReq);
-int32_t ctgPutRmTbToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, const char *tbName, bool syncReq);
-int32_t ctgPutUpdateVgToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, SDBVgInfo* dbInfo, bool syncReq);
-int32_t ctgPutUpdateTbToQueue(SCatalog* pCtg, STableMetaOutput *output, bool syncReq);
-int32_t ctgPutUpdateUserToQueue(SCatalog* pCtg, SGetUserAuthRsp *pAuth, bool syncReq);
+int32_t ctgDropDbCacheEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId);
+int32_t ctgDropStbMetaEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, const char *stbName, uint64_t suid, bool syncReq);
+int32_t ctgDropTbMetaEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, const char *tbName, bool syncReq);
+int32_t ctgUpdateVgroupEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, SDBVgInfo* dbInfo, bool syncReq);
+int32_t ctgUpdateTbMetaEnqueue(SCatalog* pCtg, STableMetaOutput *output, bool syncReq);
+int32_t ctgUpdateUserEnqueue(SCatalog* pCtg, SGetUserAuthRsp *pAuth, bool syncReq);
+int32_t ctgUpdateVgEpsetEnqueue(SCatalog* pCtg, char *dbFName, int32_t vgId, SEpSet* pEpSet);
 int32_t ctgMetaRentInit(SCtgRentMgmt *mgmt, uint32_t rentSec, int8_t type);
 int32_t ctgMetaRentAdd(SCtgRentMgmt *mgmt, void *meta, int64_t id, int32_t size);
 int32_t ctgMetaRentGet(SCtgRentMgmt *mgmt, void **res, uint32_t *num, int32_t size);

source/libs/catalog/src/catalog.c

@@ -41,9 +41,9 @@ int32_t ctgRemoveTbMetaFromCache(SCatalog* pCtg, SName* pTableName, bool syncReq
   tNameGetFullDbName(pTableName, dbFName);
   
   if (TSDB_SUPER_TABLE == tblMeta->tableType) {
-    CTG_ERR_JRET(ctgPutRmStbToQueue(pCtg, dbFName, tbCtx.tbInfo.dbId, pTableName->tname, tblMeta->suid, syncReq));
+    CTG_ERR_JRET(ctgDropStbMetaEnqueue(pCtg, dbFName, tbCtx.tbInfo.dbId, pTableName->tname, tblMeta->suid, syncReq));
   } else {
-    CTG_ERR_JRET(ctgPutRmTbToQueue(pCtg, dbFName, tbCtx.tbInfo.dbId, pTableName->tname, syncReq));
+    CTG_ERR_JRET(ctgDropTbMetaEnqueue(pCtg, dbFName, tbCtx.tbInfo.dbId, pTableName->tname, syncReq));
   }
  
 _return:
@@ -72,7 +72,7 @@ int32_t ctgGetDBVgInfo(SCatalog* pCtg, void *pTrans, const SEpSet* pMgmtEps, con
 
   CTG_ERR_JRET(ctgCloneVgInfo(DbOut.dbVgroup, pInfo));
 
-  CTG_ERR_RET(ctgPutUpdateVgToQueue(pCtg, dbFName, DbOut.dbId, DbOut.dbVgroup, false));
+  CTG_ERR_RET(ctgUpdateVgroupEnqueue(pCtg, dbFName, DbOut.dbId, DbOut.dbVgroup, false));
 
   return TSDB_CODE_SUCCESS;
 
@@ -108,13 +108,13 @@ int32_t ctgRefreshDBVgInfo(SCatalog* pCtg, void *pTrans, const SEpSet* pMgmtEps,
   if (code) {
     if (CTG_DB_NOT_EXIST(code) && (NULL != dbCache)) {
       ctgDebug("db no longer exist, dbFName:%s, dbId:%" PRIx64, input.db, input.dbId);
-      ctgPutRmDBToQueue(pCtg, input.db, input.dbId);
+      ctgDropDbCacheEnqueue(pCtg, input.db, input.dbId);
     }
 
     CTG_ERR_RET(code);
   }
 
-  CTG_ERR_RET(ctgPutUpdateVgToQueue(pCtg, dbFName, DbOut.dbId, DbOut.dbVgroup, true));
+  CTG_ERR_RET(ctgUpdateVgroupEnqueue(pCtg, dbFName, DbOut.dbId, DbOut.dbVgroup, true));
 
   return TSDB_CODE_SUCCESS;
 }
@@ -201,7 +201,7 @@ int32_t ctgRefreshTbMeta(CTG_PARAMS, SCtgTbMetaCtx* ctx, STableMetaOutput **pOut
     CTG_ERR_JRET(ctgCloneMetaOutput(output, pOutput));
   }
 
-  CTG_ERR_JRET(ctgPutUpdateTbToQueue(pCtg, output, syncReq));
+  CTG_ERR_JRET(ctgUpdateTbMetaEnqueue(pCtg, output, syncReq));
 
   return TSDB_CODE_SUCCESS;
 
@@ -298,9 +298,9 @@ _return:
     }
 
     if (TSDB_SUPER_TABLE == ctx->tbInfo.tbType) {
-      ctgPutRmStbToQueue(pCtg, dbFName, ctx->tbInfo.dbId, ctx->pName->tname, ctx->tbInfo.suid, false);
+      ctgDropStbMetaEnqueue(pCtg, dbFName, ctx->tbInfo.dbId, ctx->pName->tname, ctx->tbInfo.suid, false);
     } else {
-      ctgPutRmTbToQueue(pCtg, dbFName, ctx->tbInfo.dbId, ctx->pName->tname, false);
+      ctgDropTbMetaEnqueue(pCtg, dbFName, ctx->tbInfo.dbId, ctx->pName->tname, false);
     }
   }
 
@@ -348,7 +348,7 @@ int32_t ctgChkAuth(SCatalog* pCtg, void *pTrans, const SEpSet* pMgmtEps, const c
 
 _return:
 
-  ctgPutUpdateUserToQueue(pCtg, &authRsp, false);
+  ctgUpdateUserEnqueue(pCtg, &authRsp, false);
 
   return TSDB_CODE_SUCCESS;
 }
@@ -670,7 +670,7 @@ int32_t catalogUpdateDBVgInfo(SCatalog* pCtg, const char* dbFName, uint64_t dbId
     CTG_ERR_JRET(TSDB_CODE_CTG_INVALID_INPUT);
   }
 
-  code = ctgPutUpdateVgToQueue(pCtg, dbFName, dbId, dbInfo, false);
+  code = ctgUpdateVgroupEnqueue(pCtg, dbFName, dbId, dbInfo, false);
 
 _return:
 
@@ -691,7 +691,7 @@ int32_t catalogRemoveDB(SCatalog* pCtg, const char* dbFName, uint64_t dbId) {
     CTG_API_LEAVE(TSDB_CODE_SUCCESS);
   }
 
-  CTG_ERR_JRET(ctgPutRmDBToQueue(pCtg, dbFName, dbId));
+  CTG_ERR_JRET(ctgDropDbCacheEnqueue(pCtg, dbFName, dbId));
 
   CTG_API_LEAVE(TSDB_CODE_SUCCESS);
   
@@ -701,7 +701,19 @@ _return:
 }
 
 int32_t catalogUpdateVgEpSet(SCatalog* pCtg, const char* dbFName, int32_t vgId, SEpSet *epSet) {
-  return 0;
+  CTG_API_ENTER();
+
+  int32_t code = 0;
+  
+  if (NULL == pCtg || NULL == dbFName || NULL == epSet) {
+    CTG_API_LEAVE(TSDB_CODE_CTG_INVALID_INPUT);
+  }
+
+  CTG_ERR_JRET(ctgUpdateVgEpsetEnqueue(pCtg, (char*)dbFName, vgId, epSet));
+
+_return:
+  
+  CTG_API_LEAVE(code);
 }
 
 int32_t catalogRemoveTableMeta(SCatalog* pCtg, SName* pTableName) {
@@ -738,7 +750,7 @@ int32_t catalogRemoveStbMeta(SCatalog* pCtg, const char* dbFName, uint64_t dbId,
     CTG_API_LEAVE(TSDB_CODE_SUCCESS);
   }
 
-  CTG_ERR_JRET(ctgPutRmStbToQueue(pCtg, dbFName, dbId, stbName, suid, true));
+  CTG_ERR_JRET(ctgDropStbMetaEnqueue(pCtg, dbFName, dbId, stbName, suid, true));
 
   CTG_API_LEAVE(TSDB_CODE_SUCCESS);
   
@@ -791,7 +803,7 @@ int32_t catalogUpdateSTableMeta(SCatalog* pCtg, STableMetaRsp *rspMsg) {
   
   CTG_ERR_JRET(queryCreateTableMetaFromMsg(rspMsg, true, &output->tbMeta));
 
-  CTG_ERR_JRET(ctgPutUpdateTbToQueue(pCtg, output, false));
+  CTG_ERR_JRET(ctgUpdateTbMetaEnqueue(pCtg, output, false));
 
   CTG_API_LEAVE(code);
   
@@ -1152,7 +1164,7 @@ int32_t catalogUpdateUserAuthInfo(SCatalog* pCtg, SGetUserAuthRsp* pAuth) {
     CTG_API_LEAVE(TSDB_CODE_CTG_INVALID_INPUT);
   }
 
-  CTG_API_LEAVE(ctgPutUpdateUserToQueue(pCtg, pAuth, false));
+  CTG_API_LEAVE(ctgUpdateUserEnqueue(pCtg, pAuth, false));
 }
 
 

source/libs/catalog/src/ctgAsync.c

@@ -95,6 +95,30 @@ int32_t ctgInitGetDbCfgTask(SCtgJob *pJob, int32_t taskIdx, char *dbFName) {
   return TSDB_CODE_SUCCESS;
 }
 
+int32_t ctgInitGetDbInfoTask(SCtgJob *pJob, int32_t taskIdx, char *dbFName) {
+  SCtgTask task = {0};
+
+  task.type = CTG_TASK_GET_DB_INFO;
+  task.taskId = taskIdx;
+  task.pJob = pJob;
+
+  task.taskCtx = taosMemoryCalloc(1, sizeof(SCtgDbInfoCtx));
+  if (NULL == task.taskCtx) {
+    CTG_ERR_RET(TSDB_CODE_OUT_OF_MEMORY);
+  }
+
+  SCtgDbInfoCtx* ctx = task.taskCtx;
+  
+  memcpy(ctx->dbFName, dbFName, sizeof(ctx->dbFName));
+
+  taosArrayPush(pJob->pTasks, &task);
+
+  qDebug("QID:%" PRIx64 " task %d type %d initialized, dbFName:%s", pJob->queryId, taskIdx, task.type, dbFName);
+
+  return TSDB_CODE_SUCCESS;
+}
+
+
 int32_t ctgInitGetTbHashTask(SCtgJob *pJob, int32_t taskIdx, SName *name) {
   SCtgTask task = {0};
 
@@ -219,8 +243,9 @@ int32_t ctgInitJob(CTG_PARAMS, SCtgJob** job, uint64_t reqId, const SCatalogReq*
   int32_t dbCfgNum = (int32_t)taosArrayGetSize(pReq->pDbCfg);
   int32_t indexNum = (int32_t)taosArrayGetSize(pReq->pIndex);
   int32_t userNum = (int32_t)taosArrayGetSize(pReq->pUser);
+  int32_t dbInfoNum = (int32_t)taosArrayGetSize(pReq->pDbInfo);
 
-  int32_t taskNum = tbMetaNum + dbVgNum + udfNum + tbHashNum + qnodeNum + dbCfgNum + indexNum + userNum;
+  int32_t taskNum = tbMetaNum + dbVgNum + udfNum + tbHashNum + qnodeNum + dbCfgNum + indexNum + userNum + dbInfoNum;
   if (taskNum <= 0) {
     ctgError("empty input for job, taskNum:%d", taskNum);
     CTG_ERR_RET(TSDB_CODE_CTG_INVALID_INPUT);
@@ -249,6 +274,7 @@ int32_t ctgInitJob(CTG_PARAMS, SCtgJob** job, uint64_t reqId, const SCatalogReq*
   pJob->dbCfgNum = dbCfgNum;
   pJob->indexNum = indexNum;
   pJob->userNum = userNum;
+  pJob->dbInfoNum = dbInfoNum;
   
   pJob->pTasks = taosArrayInit(taskNum, sizeof(SCtgTask));
   
@@ -268,6 +294,11 @@ int32_t ctgInitJob(CTG_PARAMS, SCtgJob** job, uint64_t reqId, const SCatalogReq*
     CTG_ERR_JRET(ctgInitGetDbCfgTask(pJob, taskIdx++, dbFName));
   }
 
+  for (int32_t i = 0; i < dbInfoNum; ++i) {
+    char *dbFName = taosArrayGet(pReq->pDbInfo, i);
+    CTG_ERR_JRET(ctgInitGetDbInfoTask(pJob, taskIdx++, dbFName));
+  }
+
   for (int32_t i = 0; i < tbMetaNum; ++i) {
     SName *name = taosArrayGet(pReq->pTableMeta, i);
     CTG_ERR_JRET(ctgInitGetTbMetaTask(pJob, taskIdx++, name));
@@ -395,6 +426,20 @@ int32_t ctgDumpDbCfgRes(SCtgTask* pTask) {
   return TSDB_CODE_SUCCESS;
 }
 
+int32_t ctgDumpDbInfoRes(SCtgTask* pTask) {
+  SCtgJob* pJob = pTask->pJob;
+  if (NULL == pJob->jobRes.pDbInfo) {
+    pJob->jobRes.pDbInfo = taosArrayInit(pJob->dbInfoNum, sizeof(SDbInfo));
+    if (NULL == pJob->jobRes.pDbInfo) {
+      CTG_ERR_RET(TSDB_CODE_OUT_OF_MEMORY);
+    }
+  }
+
+  taosArrayPush(pJob->jobRes.pDbInfo, pTask->res);
+
+  return TSDB_CODE_SUCCESS;
+}
+
 int32_t ctgDumpUdfRes(SCtgTask* pTask) {
   SCtgJob* pJob = pTask->pJob;
   if (NULL == pJob->jobRes.pUdfList) {
@@ -620,7 +665,7 @@ int32_t ctgHandleGetDbVgRsp(SCtgTask* pTask, int32_t reqType, const SDataBuf *pM
 
       CTG_ERR_JRET(ctgGenerateVgList(pCtg, pOut->dbVgroup->vgHash, (SArray**)&pTask->res));
       
-      CTG_ERR_JRET(ctgPutUpdateVgToQueue(pCtg, ctx->dbFName, pOut->dbId, pOut->dbVgroup, false));
+      CTG_ERR_JRET(ctgUpdateVgroupEnqueue(pCtg, ctx->dbFName, pOut->dbId, pOut->dbVgroup, false));
       pOut->dbVgroup = NULL;
 
       break;
@@ -659,7 +704,7 @@ int32_t ctgHandleGetTbHashRsp(SCtgTask* pTask, int32_t reqType, const SDataBuf *
 
       CTG_ERR_JRET(ctgGetVgInfoFromHashValue(pCtg, pOut->dbVgroup, ctx->pName, (SVgroupInfo*)pTask->res));
       
-      CTG_ERR_JRET(ctgPutUpdateVgToQueue(pCtg, ctx->dbFName, pOut->dbId, pOut->dbVgroup, false));
+      CTG_ERR_JRET(ctgUpdateVgroupEnqueue(pCtg, ctx->dbFName, pOut->dbId, pOut->dbVgroup, false));
       pOut->dbVgroup = NULL;
 
       break;
@@ -691,6 +736,11 @@ _return:
   CTG_RET(code);
 }
 
+int32_t ctgHandleGetDbInfoRsp(SCtgTask* pTask, int32_t reqType, const SDataBuf *pMsg, int32_t rspCode) {
+  CTG_RET(TSDB_CODE_APP_ERROR);
+}
+
+
 int32_t ctgHandleGetQnodeRsp(SCtgTask* pTask, int32_t reqType, const SDataBuf *pMsg, int32_t rspCode) {
   int32_t code = 0;
   CTG_ERR_JRET(ctgProcessRspMsg(pTask->msgCtx.out, reqType, pMsg->pData, pMsg->len, rspCode, pTask->msgCtx.target));
@@ -769,7 +819,7 @@ _return:
     }
   }
 
-  ctgPutUpdateUserToQueue(pCtg, pOut, false);
+  ctgUpdateUserEnqueue(pCtg, pOut, false);
   taosMemoryFreeClear(pTask->msgCtx.out);
 
   ctgHandleTaskEnd(pTask, code);
@@ -933,6 +983,41 @@ int32_t ctgLaunchGetDbCfgTask(SCtgTask *pTask) {
   return TSDB_CODE_SUCCESS;
 }
 
+int32_t ctgLaunchGetDbInfoTask(SCtgTask *pTask) {
+  int32_t code = 0;
+  SCatalog* pCtg = pTask->pJob->pCtg; 
+  void *pTrans = pTask->pJob->pTrans; 
+  const SEpSet* pMgmtEps = &pTask->pJob->pMgmtEps;
+  SCtgDBCache *dbCache = NULL;
+  SCtgDbInfoCtx* pCtx = (SCtgDbInfoCtx*)pTask->taskCtx;
+
+  pTask->res = taosMemoryCalloc(1, sizeof(SDbInfo));
+  if (NULL == pTask->res) {
+    CTG_ERR_RET(TSDB_CODE_OUT_OF_MEMORY);
+  }
+
+  SDbInfo* pInfo = (SDbInfo*)pTask->res;
+  CTG_ERR_RET(ctgAcquireVgInfoFromCache(pCtg, pCtx->dbFName, &dbCache));
+  if (NULL != dbCache) {
+    pInfo->vgVer = dbCache->vgInfo->vgVersion;
+    pInfo->dbId = dbCache->dbId;
+    pInfo->tbNum = dbCache->vgInfo->numOfTable;
+  } else {
+    pInfo->vgVer = CTG_DEFAULT_INVALID_VERSION;
+  }
+
+  CTG_ERR_JRET(ctgHandleTaskEnd(pTask, 0));
+
+_return:
+
+  if (dbCache) {
+    ctgReleaseVgInfo(dbCache);
+    ctgReleaseDBCache(pCtg, dbCache);
+  }
+
+  CTG_RET(code);
+}
+
 int32_t ctgLaunchGetIndexTask(SCtgTask *pTask) {
   SCatalog* pCtg = pTask->pJob->pCtg; 
   void *pTrans = pTask->pJob->pTrans; 
@@ -992,6 +1077,7 @@ SCtgAsyncFps gCtgAsyncFps[] = {
   {ctgLaunchGetQnodeTask,  ctgHandleGetQnodeRsp,  ctgDumpQnodeRes},
   {ctgLaunchGetDbVgTask,   ctgHandleGetDbVgRsp,   ctgDumpDbVgRes},
   {ctgLaunchGetDbCfgTask,  ctgHandleGetDbCfgRsp,  ctgDumpDbCfgRes},
+  {ctgLaunchGetDbInfoTask, ctgHandleGetDbInfoRsp, ctgDumpDbInfoRes},
   {ctgLaunchGetTbMetaTask, ctgHandleGetTbMetaRsp, ctgDumpTbMetaRes},
   {ctgLaunchGetTbHashTask, ctgHandleGetTbHashRsp, ctgDumpTbHashRes},
   {ctgLaunchGetIndexTask,  ctgHandleGetIndexRsp,  ctgDumpIndexRes},

source/libs/catalog/src/ctgCache.c

@@ -19,37 +19,43 @@
 #include "catalogInt.h"
 #include "systable.h"
 
-SCtgAction gCtgAction[CTG_ACT_MAX] = {
+SCtgOperation gCtgCacheOperation[CTG_OP_MAX] = {
   {
-    CTG_ACT_UPDATE_VG,
+    CTG_OP_UPDATE_VGROUP,
     "update vgInfo",
-    ctgActUpdateVg
+    ctgOpUpdateVgroup
   },
   {
-    CTG_ACT_UPDATE_TBL,
+    CTG_OP_UPDATE_TB_META,
     "update tbMeta",
-    ctgActUpdateTb
+    ctgOpUpdateTbMeta
   },
   {
-    CTG_ACT_REMOVE_DB,
-    "remove DB",
-    ctgActRemoveDB
+    CTG_OP_DROP_DB_CACHE,
+    "drop DB",
+    ctgOpDropDbCache
   },
   {
-    CTG_ACT_REMOVE_STB,
-    "remove stbMeta",
-    ctgActRemoveStb
+    CTG_OP_DROP_STB_META,
+    "drop stbMeta",
+    ctgOpDropStbMeta
   },
   {
-    CTG_ACT_REMOVE_TBL,
-    "remove tbMeta",
-    ctgActRemoveTb
+    CTG_OP_DROP_TB_META,
+    "drop tbMeta",
+    ctgOpDropTbMeta
   },
   {
-    CTG_ACT_UPDATE_USER,
+    CTG_OP_UPDATE_USER,
     "update user",
-    ctgActUpdateUser
+    ctgOpUpdateUser
+  },
+  {
+    CTG_OP_UPDATE_VG_EPSET,
+    "update epset",
+    ctgOpUpdateEpset
   }
+  
 };
 
 
@@ -405,7 +411,7 @@ int32_t ctgReadTbVerFromCache(SCatalog *pCtg, const SName *pTableName, int32_t *
 }
 
 
-int32_t ctgGetTbTypeFromCache(SCatalog* pCtg, const char* dbFName, const char *tableName, int32_t *tbType) {
+int32_t ctgReadTbTypeFromCache(SCatalog* pCtg, const char* dbFName, const char *tableName, int32_t *tbType) {
   if (NULL == pCtg->dbCache) {
     ctgWarn("empty db cache, dbFName:%s, tbName:%s", dbFName, tableName);  
     return TSDB_CODE_SUCCESS;
@@ -491,7 +497,7 @@ _return:
 }
 
 
-void ctgWaitAction(SCtgMetaAction *action) {
+void ctgWaitOpDone(SCtgCacheOperation *action) {
   while (true) {
     tsem_wait(&gCtgMgmt.queue.rspSem);
     
@@ -509,7 +515,7 @@ void ctgWaitAction(SCtgMetaAction *action) {
   }
 }
 
-void ctgPopAction(SCtgMetaAction **action) {
+void ctgDequeue(SCtgCacheOperation **op) {
   SCtgQNode *orig = gCtgMgmt.queue.head;
   
   SCtgQNode *node = gCtgMgmt.queue.head->next;
@@ -519,20 +525,20 @@ void ctgPopAction(SCtgMetaAction **action) {
   
   taosMemoryFreeClear(orig);
 
-  *action = &node->action;
+  *op = &node->op;
 }
 
 
-int32_t ctgPushAction(SCatalog* pCtg, SCtgMetaAction *action) {
+int32_t ctgEnqueue(SCatalog* pCtg, SCtgCacheOperation *operation) {
   SCtgQNode *node = taosMemoryCalloc(1, sizeof(SCtgQNode));
   if (NULL == node) {
     qError("calloc %d failed", (int32_t)sizeof(SCtgQNode));
     CTG_RET(TSDB_CODE_CTG_MEM_ERROR);
   }
 
-  action->seqId = atomic_add_fetch_64(&gCtgMgmt.queue.seqId, 1);
+  operation->seqId = atomic_add_fetch_64(&gCtgMgmt.queue.seqId, 1);
   
-  node->action = *action;
+  node->op = *operation;
 
   CTG_LOCK(CTG_WRITE, &gCtgMgmt.queue.qlock);
   gCtgMgmt.queue.tail->next = node;
@@ -544,19 +550,19 @@ int32_t ctgPushAction(SCatalog* pCtg, SCtgMetaAction *action) {
 
   tsem_post(&gCtgMgmt.queue.reqSem);
 
-  ctgDebug("action [%s] added into queue", gCtgAction[action->act].name);
+  ctgDebug("action [%s] added into queue", gCtgCacheOperation[operation->opId].name);
 
-  if (action->syncReq) {
-    ctgWaitAction(action);
+  if (operation->syncReq) {
+    ctgWaitOpDone(operation);
   }
 
   return TSDB_CODE_SUCCESS;
 }
 
 
-int32_t ctgPutRmDBToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId) {
+int32_t ctgDropDbCacheEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId) {
   int32_t code = 0;
-  SCtgMetaAction action= {.act = CTG_ACT_REMOVE_DB};
+  SCtgCacheOperation action= {.opId = CTG_OP_DROP_DB_CACHE};
   SCtgRemoveDBMsg *msg = taosMemoryMalloc(sizeof(SCtgRemoveDBMsg));
   if (NULL == msg) {
     ctgError("malloc %d failed", (int32_t)sizeof(SCtgRemoveDBMsg));
@@ -574,7 +580,7 @@ int32_t ctgPutRmDBToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId) {
 
   action.data = msg;
 
-  CTG_ERR_JRET(ctgPushAction(pCtg, &action));
+  CTG_ERR_JRET(ctgEnqueue(pCtg, &action));
 
   return TSDB_CODE_SUCCESS;
 
@@ -585,9 +591,9 @@ _return:
 }
 
 
-int32_t ctgPutRmStbToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, const char *stbName, uint64_t suid, bool syncReq) {
+int32_t ctgDropStbMetaEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, const char *stbName, uint64_t suid, bool syncReq) {
   int32_t code = 0;
-  SCtgMetaAction action= {.act = CTG_ACT_REMOVE_STB, .syncReq = syncReq};
+  SCtgCacheOperation action= {.opId = CTG_OP_DROP_STB_META, .syncReq = syncReq};
   SCtgRemoveStbMsg *msg = taosMemoryMalloc(sizeof(SCtgRemoveStbMsg));
   if (NULL == msg) {
     ctgError("malloc %d failed", (int32_t)sizeof(SCtgRemoveStbMsg));
@@ -602,7 +608,7 @@ int32_t ctgPutRmStbToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, co
 
   action.data = msg;
 
-  CTG_ERR_JRET(ctgPushAction(pCtg, &action));
+  CTG_ERR_JRET(ctgEnqueue(pCtg, &action));
 
   return TSDB_CODE_SUCCESS;
 
@@ -614,9 +620,9 @@ _return:
 
 
 
-int32_t ctgPutRmTbToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, const char *tbName, bool syncReq) {
+int32_t ctgDropTbMetaEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, const char *tbName, bool syncReq) {
   int32_t code = 0;
-  SCtgMetaAction action= {.act = CTG_ACT_REMOVE_TBL, .syncReq = syncReq};
+  SCtgCacheOperation action= {.opId = CTG_OP_DROP_TB_META, .syncReq = syncReq};
   SCtgRemoveTblMsg *msg = taosMemoryMalloc(sizeof(SCtgRemoveTblMsg));
   if (NULL == msg) {
     ctgError("malloc %d failed", (int32_t)sizeof(SCtgRemoveTblMsg));
@@ -630,7 +636,7 @@ int32_t ctgPutRmTbToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, con
 
   action.data = msg;
 
-  CTG_ERR_JRET(ctgPushAction(pCtg, &action));
+  CTG_ERR_JRET(ctgEnqueue(pCtg, &action));
 
   return TSDB_CODE_SUCCESS;
 
@@ -640,9 +646,9 @@ _return:
   CTG_RET(code);
 }
 
-int32_t ctgPutUpdateVgToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, SDBVgInfo* dbInfo, bool syncReq) {
+int32_t ctgUpdateVgroupEnqueue(SCatalog* pCtg, const char *dbFName, int64_t dbId, SDBVgInfo* dbInfo, bool syncReq) {
   int32_t code = 0;
-  SCtgMetaAction action= {.act = CTG_ACT_UPDATE_VG, .syncReq = syncReq};
+  SCtgCacheOperation action= {.opId = CTG_OP_UPDATE_VGROUP, .syncReq = syncReq};
   SCtgUpdateVgMsg *msg = taosMemoryMalloc(sizeof(SCtgUpdateVgMsg));
   if (NULL == msg) {
     ctgError("malloc %d failed", (int32_t)sizeof(SCtgUpdateVgMsg));
@@ -662,7 +668,7 @@ int32_t ctgPutUpdateVgToQueue(SCatalog* pCtg, const char *dbFName, int64_t dbId,
 
   action.data = msg;
 
-  CTG_ERR_JRET(ctgPushAction(pCtg, &action));
+  CTG_ERR_JRET(ctgEnqueue(pCtg, &action));
 
   return TSDB_CODE_SUCCESS;
 
@@ -673,9 +679,9 @@ _return:
   CTG_RET(code);
 }
 
-int32_t ctgPutUpdateTbToQueue(SCatalog* pCtg, STableMetaOutput *output, bool syncReq) {
+int32_t ctgUpdateTbMetaEnqueue(SCatalog* pCtg, STableMetaOutput *output, bool syncReq) {
   int32_t code = 0;
-  SCtgMetaAction action= {.act = CTG_ACT_UPDATE_TBL, .syncReq = syncReq};
+  SCtgCacheOperation action= {.opId = CTG_OP_UPDATE_TB_META, .syncReq = syncReq};
   SCtgUpdateTblMsg *msg = taosMemoryMalloc(sizeof(SCtgUpdateTblMsg));
   if (NULL == msg) {
     ctgError("malloc %d failed", (int32_t)sizeof(SCtgUpdateTblMsg));
@@ -692,7 +698,34 @@ int32_t ctgPutUpdateTbToQueue(SCatalog* pCtg, STableMetaOutput *output, bool syn
 
   action.data = msg;
 
-  CTG_ERR_JRET(ctgPushAction(pCtg, &action));
+  CTG_ERR_JRET(ctgEnqueue(pCtg, &action));
+
+  return TSDB_CODE_SUCCESS;
+  
+_return:
+
+  taosMemoryFreeClear(msg);
+  
+  CTG_RET(code);
+}
+
+int32_t ctgUpdateVgEpsetEnqueue(SCatalog* pCtg, char *dbFName, int32_t vgId, SEpSet* pEpSet) {
+  int32_t code = 0;
+  SCtgCacheOperation operation= {.opId = CTG_OP_UPDATE_VG_EPSET};
+  SCtgUpdateEpsetMsg *msg = taosMemoryMalloc(sizeof(SCtgUpdateEpsetMsg));
+  if (NULL == msg) {
+    ctgError("malloc %d failed", (int32_t)sizeof(SCtgUpdateEpsetMsg));
+    CTG_ERR_RET(TSDB_CODE_CTG_MEM_ERROR);
+  }
+
+  msg->pCtg = pCtg;
+  strcpy(msg->dbFName, dbFName);
+  msg->vgId = vgId;
+  msg->epSet = *pEpSet;
+
+  operation.data = msg;
+
+  CTG_ERR_JRET(ctgEnqueue(pCtg, &operation));
 
   return TSDB_CODE_SUCCESS;
   
@@ -703,9 +736,11 @@ _return:
   CTG_RET(code);
 }
 
-int32_t ctgPutUpdateUserToQueue(SCatalog* pCtg, SGetUserAuthRsp *pAuth, bool syncReq) {
+
+
+int32_t ctgUpdateUserEnqueue(SCatalog* pCtg, SGetUserAuthRsp *pAuth, bool syncReq) {
   int32_t code = 0;
-  SCtgMetaAction action= {.act = CTG_ACT_UPDATE_USER, .syncReq = syncReq};
+  SCtgCacheOperation action= {.opId = CTG_OP_UPDATE_USER, .syncReq = syncReq};
   SCtgUpdateUserMsg *msg = taosMemoryMalloc(sizeof(SCtgUpdateUserMsg));
   if (NULL == msg) {
     ctgError("malloc %d failed", (int32_t)sizeof(SCtgUpdateUserMsg));
@@ -717,7 +752,7 @@ int32_t ctgPutUpdateUserToQueue(SCatalog* pCtg, SGetUserAuthRsp *pAuth, bool syn
 
   action.data = msg;
 
-  CTG_ERR_JRET(ctgPushAction(pCtg, &action));
+  CTG_ERR_JRET(ctgEnqueue(pCtg, &action));
   
   return TSDB_CODE_SUCCESS;
   
@@ -1219,7 +1254,7 @@ int32_t ctgUpdateTbMetaToCache(SCatalog* pCtg, STableMetaOutput* pOut, bool sync
   int32_t code = 0;
   
   CTG_ERR_RET(ctgCloneMetaOutput(pOut, &pOutput));
-  CTG_ERR_JRET(ctgPutUpdateTbToQueue(pCtg, pOutput, syncReq));
+  CTG_ERR_JRET(ctgUpdateTbMetaEnqueue(pCtg, pOutput, syncReq));
 
   return TSDB_CODE_SUCCESS;
   
@@ -1230,9 +1265,9 @@ _return:
 }
 
 
-int32_t ctgActUpdateVg(SCtgMetaAction *action) {
+int32_t ctgOpUpdateVgroup(SCtgCacheOperation *operation) {
   int32_t code = 0;
-  SCtgUpdateVgMsg *msg = action->data;
+  SCtgUpdateVgMsg *msg = operation->data;
   
   CTG_ERR_JRET(ctgWriteDBVgInfoToCache(msg->pCtg, msg->dbFName, msg->dbId, &msg->dbInfo));
 
@@ -1244,9 +1279,9 @@ _return:
   CTG_RET(code);
 }
 
-int32_t ctgActRemoveDB(SCtgMetaAction *action) {
+int32_t ctgOpDropDbCache(SCtgCacheOperation *operation) {
   int32_t code = 0;
-  SCtgRemoveDBMsg *msg = action->data;
+  SCtgRemoveDBMsg *msg = operation->data;
   SCatalog* pCtg = msg->pCtg;
 
   SCtgDBCache *dbCache = NULL;
@@ -1270,9 +1305,9 @@ _return:
 }
 
 
-int32_t ctgActUpdateTb(SCtgMetaAction *action) {
+int32_t ctgOpUpdateTbMeta(SCtgCacheOperation *operation) {
   int32_t code = 0;
-  SCtgUpdateTblMsg *msg = action->data;
+  SCtgUpdateTblMsg *msg = operation->data;
   SCatalog* pCtg = msg->pCtg;
   STableMetaOutput* output = msg->output;
   SCtgDBCache *dbCache = NULL;
@@ -1316,9 +1351,9 @@ _return:
 }
 
 
-int32_t ctgActRemoveStb(SCtgMetaAction *action) {
+int32_t ctgOpDropStbMeta(SCtgCacheOperation *operation) {
   int32_t code = 0;
-  SCtgRemoveStbMsg *msg = action->data;
+  SCtgRemoveStbMsg *msg = operation->data;
   SCatalog* pCtg = msg->pCtg;
 
   SCtgDBCache *dbCache = NULL;
@@ -1362,9 +1397,9 @@ _return:
   CTG_RET(code);
 }
 
-int32_t ctgActRemoveTb(SCtgMetaAction *action) {
+int32_t ctgOpDropTbMeta(SCtgCacheOperation *operation) {
   int32_t code = 0;
-  SCtgRemoveTblMsg *msg = action->data;
+  SCtgRemoveTblMsg *msg = operation->data;
   SCatalog* pCtg = msg->pCtg;
 
   SCtgDBCache *dbCache = NULL;
@@ -1397,9 +1432,9 @@ _return:
   CTG_RET(code);
 }
 
-int32_t ctgActUpdateUser(SCtgMetaAction *action) {
+int32_t ctgOpUpdateUser(SCtgCacheOperation *operation) {
   int32_t code = 0;
-  SCtgUpdateUserMsg *msg = action->data;
+  SCtgUpdateUserMsg *msg = operation->data;
   SCatalog* pCtg = msg->pCtg;
 
   if (NULL == pCtg->userCache) {
@@ -1460,14 +1495,60 @@ _return:
   CTG_RET(code);
 }
 
-void ctgUpdateThreadFuncUnexpectedStopped(void) {
+int32_t ctgOpUpdateEpset(SCtgCacheOperation *operation) {
+  int32_t code = 0;
+  SCtgUpdateEpsetMsg *msg = operation->data;
+  SCatalog* pCtg = msg->pCtg;
+  
+  SCtgDBCache *dbCache = NULL;
+  CTG_ERR_RET(ctgAcquireDBCache(pCtg, msg->dbFName, &dbCache));
+  if (NULL == dbCache) {
+    ctgDebug("db %s not exist, ignore epset update", msg->dbFName);
+    goto _return;
+  }
+
+  SDBVgInfo *vgInfo = NULL;
+  CTG_ERR_RET(ctgWAcquireVgInfo(pCtg, dbCache));
+  
+  if (NULL == dbCache->vgInfo) {
+    ctgWReleaseVgInfo(dbCache);
+    ctgDebug("vgroup in db %s not cached, ignore epset update", msg->dbFName);
+    goto _return;
+  }
+  
+  SVgroupInfo* pInfo = taosHashGet(dbCache->vgInfo->vgHash, &msg->vgId, sizeof(msg->vgId));
+  if (NULL == pInfo) {
+    ctgWReleaseVgInfo(dbCache);
+    ctgDebug("no vgroup %d in db %s, ignore epset update", msg->vgId, msg->dbFName);
+    goto _return;
+  }
+
+  pInfo->epSet = msg->epSet;
+
+  ctgDebug("epset in vgroup %d updated, dbFName:%s", pInfo->vgId, msg->dbFName);
+
+  ctgWReleaseVgInfo(dbCache);
+
+_return:
+
+  if (dbCache) {
+    ctgReleaseDBCache(msg->pCtg, dbCache);
+  }
+
+  taosMemoryFreeClear(msg);
+  
+  CTG_RET(code);
+}
+
+
+void ctgUpdateThreadUnexpectedStopped(void) {
   if (CTG_IS_LOCKED(&gCtgMgmt.lock) > 0) CTG_UNLOCK(CTG_READ, &gCtgMgmt.lock);
 }
 
 void* ctgUpdateThreadFunc(void* param) {
   setThreadName("catalog");
 #ifdef WINDOWS
-  atexit(ctgUpdateThreadFuncUnexpectedStopped);
+  atexit(ctgUpdateThreadUnexpectedStopped);
 #endif
   qInfo("catalog update thread started");
 
@@ -1483,17 +1564,17 @@ void* ctgUpdateThreadFunc(void* param) {
       break;
     }
 
-    SCtgMetaAction *action = NULL;
-    ctgPopAction(&action);
-    SCatalog *pCtg = ((SCtgUpdateMsgHeader *)action->data)->pCtg;
+    SCtgCacheOperation *operation = NULL;
+    ctgDequeue(&operation);
+    SCatalog *pCtg = ((SCtgUpdateMsgHeader *)operation->data)->pCtg;
 
-    ctgDebug("process [%s] action", gCtgAction[action->act].name);
+    ctgDebug("process [%s] operation", gCtgCacheOperation[operation->opId].name);
     
-    (*gCtgAction[action->act].func)(action);
+    (*gCtgCacheOperation[operation->opId].func)(operation);
 
-    gCtgMgmt.queue.seqDone = action->seqId;
+    gCtgMgmt.queue.seqDone = operation->seqId;
 
-    if (action->syncReq) {
+    if (operation->syncReq) {
       tsem_post(&gCtgMgmt.queue.rspSem);
     }
 

source/libs/catalog/src/ctgUtil.c

@@ -42,6 +42,9 @@ void ctgFreeSMetaData(SMetaData* pData) {
   }
   taosArrayDestroy(pData->pDbCfg);
   pData->pDbCfg = NULL;
+
+  taosArrayDestroy(pData->pDbInfo);
+  pData->pDbInfo = NULL;
   
   taosArrayDestroy(pData->pIndex);
   pData->pIndex = NULL;
@@ -293,9 +296,12 @@ void ctgFreeTask(SCtgTask* pTask) {
     }
     case CTG_TASK_GET_DB_CFG: {
       taosMemoryFreeClear(pTask->taskCtx);      
-      if (pTask->res) {
-        taosMemoryFreeClear(pTask->res);
-      }
+      taosMemoryFreeClear(pTask->res);
+      break;
+    }
+    case CTG_TASK_GET_DB_INFO: {
+      taosMemoryFreeClear(pTask->taskCtx);      
+      taosMemoryFreeClear(pTask->res);
       break;
     }
     case CTG_TASK_GET_TB_HASH: {

source/libs/catalog/test/catalogTests.cpp

@@ -41,7 +41,7 @@
 namespace {
 
 extern "C" int32_t ctgdGetClusterCacheNum(struct SCatalog* pCatalog, int32_t type);
-extern "C" int32_t ctgActUpdateTb(SCtgMetaAction *action);
+extern "C" int32_t ctgOpUpdateTbMeta(SCtgCacheOperation *action);
 extern "C" int32_t ctgdEnableDebug(char *option);
 extern "C" int32_t ctgdGetStatNum(char *option, void *res);
 
@@ -888,9 +888,9 @@ void *ctgTestSetCtableMetaThread(void *param) {
   int32_t          n = 0;
   STableMetaOutput *output = NULL;
 
-  SCtgMetaAction action = {0};
+  SCtgCacheOperation operation = {0};
   
-  action.act = CTG_ACT_UPDATE_TBL;
+  operation.opId = CTG_OP_UPDATE_TB_META;
 
   while (!ctgTestStop) {
     output = (STableMetaOutput *)taosMemoryMalloc(sizeof(STableMetaOutput));
@@ -899,9 +899,9 @@ void *ctgTestSetCtableMetaThread(void *param) {
     SCtgUpdateTblMsg *msg = (SCtgUpdateTblMsg *)taosMemoryMalloc(sizeof(SCtgUpdateTblMsg));
     msg->pCtg = pCtg;
     msg->output = output;
-    action.data = msg;
+    operation.data = msg;
 
-    code = ctgActUpdateTb(&action);
+    code = ctgOpUpdateTbMeta(&operation);
     if (code) {
       assert(0);
     }

source/libs/executor/inc/executorimpl.h

@@ -334,6 +334,8 @@ typedef struct STableScanInfo {
   int32_t         dataBlockLoadFlag;
   double          sampleRatio;  // data block sample ratio, 1 by default
   SInterval       interval;     // if the upstream is an interval operator, the interval info is also kept here to get the time window to check if current data block needs to be loaded.
+
+  int32_t         curTWinIdx;
 } STableScanInfo;
 
 typedef struct STagScanInfo {
@@ -373,30 +375,33 @@ typedef struct SessionWindowSupporter {
 } SessionWindowSupporter;
 
 typedef struct SStreamBlockScanInfo {
-  SArray*      pBlockLists;      // multiple SSDatablock.
-  SSDataBlock* pRes;             // result SSDataBlock
-  SSDataBlock* pUpdateRes;       // update SSDataBlock
-  int32_t      updateResIndex;
-  int32_t      blockType;        // current block type
-  int32_t      validBlockIndex;  // Is current data has returned?
-  SColumnInfo* pCols;            // the output column info
-  uint64_t     numOfExec;        // execution times
-  void*        streamBlockReader;// stream block reader handle
-  SArray*      pColMatchInfo;    //
-  SNode*       pCondition;
-  SArray*      tsArray;
-  SUpdateInfo* pUpdateInfo;
-  int32_t      primaryTsIndex;    // primary time stamp slot id
-  void*        pDataReader;
-  SReadHandle  readHandle;
-  uint64_t     tableUid;         // queried super table uid
+  SArray*         pBlockLists;      // multiple SSDatablock.
+  SSDataBlock*    pRes;             // result SSDataBlock
+  SSDataBlock*    pUpdateRes;       // update SSDataBlock
+  int32_t         updateResIndex;
+  int32_t         blockType;        // current block type
+  int32_t         validBlockIndex;  // Is current data has returned?
+  SColumnInfo*    pCols;            // the output column info
+  uint64_t        numOfExec;        // execution times
+  void*           streamBlockReader;// stream block reader handle
+  SArray*         pColMatchInfo;    //
+  SNode*          pCondition;
+  SArray*         tsArray;
+  SUpdateInfo*    pUpdateInfo;
+
+  SExprInfo*      pPseudoExpr;
+  int32_t         numOfPseudoExpr;
+
+  int32_t         primaryTsIndex;    // primary time stamp slot id
+  void*           pDataReader;
+  SReadHandle     readHandle;
+  uint64_t        tableUid;         // queried super table uid
   EStreamScanMode scanMode;
   SOperatorInfo* pOperatorDumy;
   SInterval      interval;     // if the upstream is an interval operator, the interval info is also kept here.
-  SCatchSupporter childAggSup;
-  SArray*      childIds;
+  SArray*        childIds;
   SessionWindowSupporter sessionSup;
-  bool         assignBlockUid; // assign block uid to groupId, temporarily used for generating rollup SMA.
+  bool            assignBlockUid; // assign block uid to groupId, temporarily used for generating rollup SMA.
 } SStreamBlockScanInfo;
 
 typedef struct SSysTableScanInfo {
@@ -435,6 +440,7 @@ typedef struct SAggSupporter {
 typedef struct STimeWindowSupp {
   int8_t           calTrigger;
   int64_t          waterMark;
+  TSKEY            maxTs;
   SColumnInfoData  timeWindowData;     // query time window info for scalar function execution.
 } STimeWindowAggSupp;
 
@@ -566,6 +572,7 @@ typedef struct SResultWindowInfo {
   SResultRowPosition pos;
   STimeWindow win;
   bool isOutput;
+  bool isClosed;
 } SResultWindowInfo;
 
 typedef struct SStreamSessionAggOperatorInfo {
@@ -735,10 +742,11 @@ SOperatorInfo* createGroupOperatorInfo(SOperatorInfo* downstream, SExprInfo* pEx
                                        SSDataBlock* pResultBlock, SArray* pGroupColList, SNode* pCondition,
                                        SExprInfo* pScalarExprInfo, int32_t numOfScalarExpr, SExecTaskInfo* pTaskInfo);
 SOperatorInfo* createDataBlockInfoScanOperator(void* dataReader, SExecTaskInfo* pTaskInfo);
-SOperatorInfo* createStreamScanOperatorInfo(void* streamReadHandle, void* pDataReader, SReadHandle* pHandle,
-                                            uint64_t uid, SSDataBlock* pResBlock, SArray* pColList,
-                                            SArray* pTableIdList, SExecTaskInfo* pTaskInfo, SNode* pCondition,
-                                            SOperatorInfo* pOperatorDumy);
+
+SOperatorInfo* createStreamScanOperatorInfo(void* pDataReader, SReadHandle* pHandle,
+    SArray* pTableIdList, STableScanPhysiNode* pTableScanNode, SExecTaskInfo* pTaskInfo,
+    STimeWindowAggSupp* pTwSup, int16_t tsColId);
+
 
 SOperatorInfo* createFillOperatorInfo(SOperatorInfo* downstream, SExprInfo* pExpr, int32_t numOfCols,
                                       SInterval* pInterval, STimeWindow* pWindow, SSDataBlock* pResBlock, int32_t fillType, SNodeListNode* fillVal,
@@ -794,8 +802,6 @@ int32_t getNumOfRowsInTimeWindow(SDataBlockInfo* pDataBlockInfo, TSKEY* pPrimary
     int32_t startPos, TSKEY ekey, __block_search_fn_t searchFn, STableQueryInfo* item,
     int32_t order);
 int32_t binarySearchForKey(char* pValue, int num, TSKEY key, int order);
-int32_t initCacheSupporter(SCatchSupporter* pCatchSup, size_t rowSize, const char* pKey,
-    const char* pDir);
 int32_t initStreamAggSupporter(SStreamAggSupporter* pSup, const char* pKey);
 SResultRow* getNewResultRow_rv(SDiskbasedBuf* pResultBuf, int64_t tableGroupId, int32_t interBufSize);
 SResultWindowInfo* getSessionTimeWindow(SArray* pWinInfos, TSKEY ts, int64_t gap,
@@ -803,6 +809,8 @@ SResultWindowInfo* getSessionTimeWindow(SArray* pWinInfos, TSKEY ts, int64_t gap
 int32_t updateSessionWindowInfo(SResultWindowInfo* pWinInfo, TSKEY* pTs, int32_t rows,
     int32_t start, int64_t gap, SHashObj* pStDeleted);
 bool functionNeedToExecute(SqlFunctionCtx* pCtx);
+
+int32_t compareTimeWindow(const void* p1, const void* p2, const void* param);
 #ifdef __cplusplus
 }
 #endif

source/libs/executor/src/executil.c

@@ -233,7 +233,7 @@ void initGroupedResultInfo(SGroupResInfo* pGroupResInfo, SHashObj* pHashmap, int
 
 void initMultiResInfoFromArrayList(SGroupResInfo* pGroupResInfo, SArray* pArrayList) {
   if (pGroupResInfo->pRows != NULL) {
-    taosArrayDestroy(pGroupResInfo->pRows);
+    taosArrayDestroyP(pGroupResInfo->pRows, taosMemoryFree);
   }
 
   pGroupResInfo->pRows = pArrayList;

source/libs/executor/src/executorimpl.c

@@ -28,13 +28,13 @@
 #include "ttime.h"
 
 #include "executorimpl.h"
+#include "index.h"
 #include "query.h"
 #include "tcompare.h"
 #include "tcompression.h"
 #include "thash.h"
 #include "ttypes.h"
 #include "vnode.h"
-#include "index.h"
 
 #define IS_MAIN_SCAN(runtime)          ((runtime)->scanFlag == MAIN_SCAN)
 #define IS_REVERSE_SCAN(runtime)       ((runtime)->scanFlag == REVERSE_SCAN)
@@ -87,7 +87,7 @@ static UNUSED_FUNC void* u_realloc(void* p, size_t __size) {
 #define realloc u_realloc
 #endif
 
-#define CLEAR_QUERY_STATUS(q, st)   ((q)->status &= (~(st)))
+#define CLEAR_QUERY_STATUS(q, st) ((q)->status &= (~(st)))
 //#define GET_NUM_OF_TABLEGROUP(q)    taosArrayGetSize((q)->tableqinfoGroupInfo.pGroupList)
 #define QUERY_IS_INTERVAL_QUERY(_q) ((_q)->interval.interval > 0)
 
@@ -3960,11 +3960,11 @@ int32_t doInitAggInfoSup(SAggSupporter* pAggSup, SqlFunctionCtx* pCtx, int32_t n
     return TSDB_CODE_OUT_OF_MEMORY;
   }
 
-  uint32_t defaultPgsz  = 0;
+  uint32_t defaultPgsz = 0;
   uint32_t defaultBufsz = 0;
   getBufferPgSize(pAggSup->resultRowSize, &defaultPgsz, &defaultBufsz);
 
-  int32_t  code = createDiskbasedBuf(&pAggSup->pResultBuf, defaultPgsz, defaultBufsz, pKey, TD_TMP_DIR_PATH);
+  int32_t code = createDiskbasedBuf(&pAggSup->pResultBuf, defaultPgsz, defaultBufsz, pKey, TD_TMP_DIR_PATH);
   if (code != TSDB_CODE_SUCCESS) {
     return code;
   }
@@ -4001,7 +4001,7 @@ void initResultSizeInfo(SOperatorInfo* pOperator, int32_t numOfRows) {
   }
 }
 
-//static STableQueryInfo* initTableQueryInfo(const STableListInfo* pTableListInfo) {
+// static STableQueryInfo* initTableQueryInfo(const STableListInfo* pTableListInfo) {
 //  int32_t size = taosArrayGetSize(pTableListInfo->pTableList);
 //  if (size == 0) {
 //    return NULL;
@@ -4434,9 +4434,11 @@ static SExecTaskInfo* createExecTaskInfo(uint64_t queryId, uint64_t taskId, EOPT
 }
 
 static tsdbReaderT doCreateDataReader(STableScanPhysiNode* pTableScanNode, SReadHandle* pHandle,
-                                      STableListInfo* pTableGroupInfo, uint64_t queryId, uint64_t taskId, SNode* pTagCond);
+                                      STableListInfo* pTableGroupInfo, uint64_t queryId, uint64_t taskId,
+                                      SNode* pTagCond);
 
-static int32_t getTableList(void* metaHandle, int32_t tableType, uint64_t tableUid, STableListInfo* pListInfo, SNode* pTagCond);
+static int32_t getTableList(void* metaHandle, int32_t tableType, uint64_t tableUid, STableListInfo* pListInfo,
+                            SNode* pTagCond);
 static SArray* extractTableIdList(const STableListInfo* pTableGroupInfo);
 static SArray* extractColumnInfo(SNodeList* pNodeList);
 
@@ -4473,7 +4475,8 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
     if (QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN == type) {
       STableScanPhysiNode* pTableScanNode = (STableScanPhysiNode*)pPhyNode;
 
-      tsdbReaderT pDataReader = doCreateDataReader(pTableScanNode, pHandle, pTableListInfo, (uint64_t)queryId, taskId, pTagCond);
+      tsdbReaderT pDataReader =
+          doCreateDataReader(pTableScanNode, pHandle, pTableListInfo, (uint64_t)queryId, taskId, pTagCond);
       if (pDataReader == NULL && terrno != 0) {
         return NULL;
       }
@@ -4492,9 +4495,8 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
     } else if (QUERY_NODE_PHYSICAL_PLAN_STREAM_SCAN == type) {
       SScanPhysiNode*      pScanPhyNode = (SScanPhysiNode*)pPhyNode;  // simple child table.
       STableScanPhysiNode* pTableScanNode = (STableScanPhysiNode*)pPhyNode;
-
-      int32_t numOfCols = 0;
-
+      STimeWindowAggSupp twSup = {.waterMark = pTableScanNode->watermark,
+          .calTrigger = pTableScanNode->triggerType, .maxTs = INT64_MIN};
       tsdbReaderT pDataReader = NULL;
       if (pHandle->vnode) {
         pDataReader = doCreateDataReader(pTableScanNode, pHandle, pTableListInfo, (uint64_t)queryId, taskId, pTagCond);
@@ -4503,24 +4505,16 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
       }
 
       if (pDataReader == NULL && terrno != 0) {
-        qDebug("pDataReader is     NULL");
+        qDebug("%s pDataReader is NULL", GET_TASKID(pTaskInfo));
         // return NULL;
       } else {
-        qDebug("pDataReader is not NULL");
+        qDebug("%s pDataReader is not NULL", GET_TASKID(pTaskInfo));
       }
 
-      SDataBlockDescNode* pDescNode = pScanPhyNode->node.pOutputDataBlockDesc;
-      SOperatorInfo*      pOperatorDumy = createTableScanOperatorInfo(pTableScanNode, pDataReader, pHandle, pTaskInfo);
-
       SArray* tableIdList = extractTableIdList(pTableListInfo);
+      SOperatorInfo* pOperator = createStreamScanOperatorInfo(pDataReader, pHandle,
+          tableIdList, pTableScanNode, pTaskInfo, &twSup, pTableScanNode->tsColId);
 
-      SSDataBlock* pResBlock = createResDataBlock(pDescNode);
-      SArray*      pCols =
-          extractColMatchInfo(pScanPhyNode->pScanCols, pDescNode, &numOfCols, pTaskInfo, COL_MATCH_FROM_COL_ID);
-
-      SOperatorInfo* pOperator =
-          createStreamScanOperatorInfo(pHandle->reader, pDataReader, pHandle, pScanPhyNode->uid, pResBlock, pCols,
-                                       tableIdList, pTaskInfo, pScanPhyNode->node.pConditions, pOperatorDumy);
       taosArrayDestroy(tableIdList);
       return pOperator;
     } else if (QUERY_NODE_PHYSICAL_PLAN_SYSTABLE_SCAN == type) {
@@ -4602,8 +4596,8 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
       pOptr = createGroupOperatorInfo(ops[0], pExprInfo, num, pResBlock, pColList, pAggNode->node.pConditions,
                                       pScalarExprInfo, numOfScalarExpr, pTaskInfo);
     } else {
-      pOptr = createAggregateOperatorInfo(ops[0], pExprInfo, num, pResBlock, pScalarExprInfo, numOfScalarExpr,
-                                          pTaskInfo);
+      pOptr =
+          createAggregateOperatorInfo(ops[0], pExprInfo, num, pResBlock, pScalarExprInfo, numOfScalarExpr, pTaskInfo);
     }
   } else if (QUERY_NODE_PHYSICAL_PLAN_INTERVAL == type || QUERY_NODE_PHYSICAL_PLAN_STREAM_INTERVAL == type) {
     SIntervalPhysiNode* pIntervalPhyNode = (SIntervalPhysiNode*)pPhyNode;
@@ -4619,7 +4613,8 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
                           .precision = ((SColumnNode*)pIntervalPhyNode->window.pTspk)->node.resType.precision};
 
     STimeWindowAggSupp as = {.waterMark = pIntervalPhyNode->window.watermark,
-                             .calTrigger = pIntervalPhyNode->window.triggerType};
+                             .calTrigger = pIntervalPhyNode->window.triggerType,
+                             .maxTs = INT64_MIN};
 
     int32_t tsSlotId = ((SColumnNode*)pIntervalPhyNode->window.pTspk)->slotId;
     pOptr = createIntervalOperatorInfo(ops[0], pExprInfo, num, pResBlock, &interval, tsSlotId, &as, pTaskInfo);
@@ -4706,6 +4701,18 @@ SOperatorInfo* createOperatorTree(SPhysiNode* pPhyNode, SExecTaskInfo* pTaskInfo
   return pOptr;
 }
 
+int32_t compareTimeWindow(const void* p1, const void* p2, const void* param) {
+  const SQueryTableDataCond *pCond = param;
+  const STimeWindow *pWin1 = p1;
+  const STimeWindow *pWin2 = p2;
+  if (pCond->order == TSDB_ORDER_ASC) {
+    return pWin1->skey - pWin2->skey;
+  } else if (pCond->order == TSDB_ORDER_DESC) {
+    return pWin2->skey - pWin1->skey;
+  }
+  return 0;
+}
+
 int32_t initQueryTableDataCond(SQueryTableDataCond* pCond, const STableScanPhysiNode* pTableScanNode) {
   pCond->loadExternalRows = false;
 
@@ -4717,16 +4724,34 @@ int32_t initQueryTableDataCond(SQueryTableDataCond* pCond, const STableScanPhysi
     return terrno;
   }
 
-  pCond->twindow = pTableScanNode->scanRange;
+  //pCond->twindow = pTableScanNode->scanRange;
+  //TODO: get it from stable scan node
+  pCond->numOfTWindows = 1;
+  pCond->twindows = taosMemoryCalloc(pCond->numOfTWindows, sizeof(STimeWindow));
+  pCond->twindows[0] = pTableScanNode->scanRange;
 
 #if 1
   // todo work around a problem, remove it later
-  if ((pCond->order == TSDB_ORDER_ASC && pCond->twindow.skey > pCond->twindow.ekey) ||
-      (pCond->order == TSDB_ORDER_DESC && pCond->twindow.skey < pCond->twindow.ekey)) {
-    TSWAP(pCond->twindow.skey, pCond->twindow.ekey);
+  for (int32_t i = 0; i < pCond->numOfTWindows; ++i) {
+    if ((pCond->order == TSDB_ORDER_ASC && pCond->twindows[i].skey > pCond->twindows[i].ekey) ||
+        (pCond->order == TSDB_ORDER_DESC && pCond->twindows[i].skey < pCond->twindows[i].ekey)) {
+      TSWAP(pCond->twindows[i].skey, pCond->twindows[i].ekey);
+    }
   }
 #endif
 
+  for (int32_t i = 0; i < pCond->numOfTWindows; ++i) {
+    if ((pCond->order == TSDB_ORDER_ASC && pCond->twindows[i].skey > pCond->twindows[i].ekey) ||
+        (pCond->order == TSDB_ORDER_DESC && pCond->twindows[i].skey < pCond->twindows[i].ekey)) {
+      TSWAP(pCond->twindows[i].skey, pCond->twindows[i].ekey);
+    }
+  }
+  taosqsort(pCond->twindows,
+            pCond->numOfTWindows,
+            sizeof(STimeWindow),
+            pCond,
+            compareTimeWindow);
+
   pCond->type = BLOCK_LOAD_OFFSET_SEQ_ORDER;
   //  pCond->type = pTableScanNode->scanFlag;
 
@@ -4885,27 +4910,30 @@ SArray* extractColMatchInfo(SNodeList* pNodeList, SDataBlockDescNode* pOutputNod
   return pList;
 }
 
-int32_t getTableList(void* metaHandle, int32_t tableType, uint64_t tableUid,
-                           STableListInfo* pListInfo, SNode* pTagCond) {
+int32_t getTableList(void* metaHandle, int32_t tableType, uint64_t tableUid, STableListInfo* pListInfo, SNode* pTagCond) {
   int32_t code = TSDB_CODE_SUCCESS;
   pListInfo->pTableList = taosArrayInit(8, sizeof(STableKeyInfo));
 
   if (tableType == TSDB_SUPER_TABLE) {
-    if(pTagCond){
+    if (pTagCond) {
+      SIndexMetaArg metaArg = {.metaHandle = tsdbGetIdx(metaHandle), .suid = tableUid};
+
       SArray* res = taosArrayInit(8, sizeof(uint64_t));
-      code = doFilterTag(pTagCond, res);
+      code = doFilterTag(pTagCond, &metaArg, res);
       if (code != TSDB_CODE_SUCCESS) {
         qError("doFilterTag error:%d", code);
         taosArrayDestroy(res);
         terrno = code;
         return code;
+      } else {
+        qDebug("doFilterTag error:%d, suid: %" PRIu64 "", code, tableUid);
       }
-      for(int i = 0; i < taosArrayGetSize(res); i++){
+      for (int i = 0; i < taosArrayGetSize(res); i++) {
         STableKeyInfo info = {.lastKey = TSKEY_INITIAL_VAL, .uid = *(uint64_t*)taosArrayGet(res, i)};
         taosArrayPush(pListInfo->pTableList, &info);
       }
       taosArrayDestroy(res);
-    }else{
+    } else {
       code = tsdbGetAllTableList(metaHandle, tableUid, pListInfo->pTableList);
     }
   } else {  // Create one table group.
@@ -4930,7 +4958,8 @@ SArray* extractTableIdList(const STableListInfo* pTableGroupInfo) {
 
 tsdbReaderT doCreateDataReader(STableScanPhysiNode* pTableScanNode, SReadHandle* pHandle,
                                STableListInfo* pTableListInfo, uint64_t queryId, uint64_t taskId, SNode* pTagCond) {
-  int32_t  code = getTableList(pHandle->meta, pTableScanNode->scan.tableType, pTableScanNode->scan.uid, pTableListInfo, pTagCond);
+  int32_t code =
+      getTableList(pHandle->meta, pTableScanNode->scan.tableType, pTableScanNode->scan.uid, pTableListInfo, pTagCond);
   if (code != TSDB_CODE_SUCCESS) {
     goto _error;
   }
@@ -4965,8 +4994,8 @@ int32_t createExecTaskInfoImpl(SSubplan* pPlan, SExecTaskInfo** pTaskInfo, SRead
     goto _complete;
   }
 
-  (*pTaskInfo)->pRoot =
-      createOperatorTree(pPlan->pNode, *pTaskInfo, pHandle, queryId, taskId, &(*pTaskInfo)->tableqinfoList, pPlan->pTagCond);
+  (*pTaskInfo)->pRoot = createOperatorTree(pPlan->pNode, *pTaskInfo, pHandle, queryId, taskId,
+                                           &(*pTaskInfo)->tableqinfoList, pPlan->pTagCond);
   if (NULL == (*pTaskInfo)->pRoot) {
     code = terrno;
     goto _complete;
@@ -5151,20 +5180,6 @@ int32_t getOperatorExplainExecInfo(SOperatorInfo* operatorInfo, SExplainExecInfo
   return TSDB_CODE_SUCCESS;
 }
 
-int32_t initCacheSupporter(SCatchSupporter* pCatchSup, size_t rowSize, const char* pKey, const char* pDir) {
-  pCatchSup->keySize = sizeof(int64_t) + sizeof(int64_t) + sizeof(TSKEY);
-  pCatchSup->pKeyBuf = taosMemoryCalloc(1, pCatchSup->keySize);
-  _hash_fn_t hashFn = taosGetDefaultHashFunction(TSDB_DATA_TYPE_BINARY);
-  pCatchSup->pWindowHashTable = taosHashInit(10000, hashFn, true, HASH_NO_LOCK);
-  if (pCatchSup->pKeyBuf == NULL || pCatchSup->pWindowHashTable == NULL) {
-    return TSDB_CODE_OUT_OF_MEMORY;
-  }
-
-  int32_t pageSize = rowSize * 32;
-  int32_t bufSize = pageSize * 4096;
-  return createDiskbasedBuf(&pCatchSup->pDataBuf, pageSize, bufSize, pKey, pDir);
-}
-
 int32_t initStreamAggSupporter(SStreamAggSupporter* pSup, const char* pKey) {
   pSup->keySize = sizeof(int64_t) + sizeof(TSKEY);
   pSup->pKeyBuf = taosMemoryCalloc(1, pSup->keySize);

source/libs/executor/src/scanoperator.c

@@ -158,7 +158,7 @@ static bool overlapWithTimeWindow(SInterval* pInterval, SDataBlockInfo* pBlockIn
   return false;
 }
 
-static void addTagPseudoColumnData(STableScanInfo* pTableScanInfo, SSDataBlock* pBlock);
+static void addTagPseudoColumnData(SReadHandle *pHandle, SExprInfo* pPseudoExpr, int32_t numOfPseudoExpr, SSDataBlock* pBlock);
 
 static int32_t loadDataBlock(SOperatorInfo* pOperator, STableScanInfo* pTableScanInfo, SSDataBlock* pBlock,
                              uint32_t* status) {
@@ -250,7 +250,7 @@ static int32_t loadDataBlock(SOperatorInfo* pOperator, STableScanInfo* pTableSca
 
   // currently only the tbname pseudo column
   if (pTableScanInfo->numOfPseudoExpr > 0) {
-    addTagPseudoColumnData(pTableScanInfo, pBlock);
+    addTagPseudoColumnData(&pTableScanInfo->readHandle, pTableScanInfo->pPseudoExpr, pTableScanInfo->numOfPseudoExpr, pBlock);
   }
 
   int64_t st = taosGetTimestampMs();
@@ -274,23 +274,31 @@ static void prepareForDescendingScan(STableScanInfo* pTableScanInfo, SqlFunction
   switchCtxOrder(pCtx, numOfOutput);
   //  setupQueryRangeForReverseScan(pTableScanInfo);
 
-  STimeWindow* pTWindow = &pTableScanInfo->cond.twindow;
-  TSWAP(pTWindow->skey, pTWindow->ekey);
   pTableScanInfo->cond.order = TSDB_ORDER_DESC;
+  for (int32_t i = 0; i < pTableScanInfo->cond.numOfTWindows; ++i) {
+    STimeWindow* pTWindow = &pTableScanInfo->cond.twindows[i];
+    TSWAP(pTWindow->skey, pTWindow->ekey);
+  }
+  SQueryTableDataCond *pCond = &pTableScanInfo->cond;
+  taosqsort(pCond->twindows,
+            pCond->numOfTWindows,
+            sizeof(STimeWindow),
+            pCond,
+            compareTimeWindow);
 }
 
-void addTagPseudoColumnData(STableScanInfo* pTableScanInfo, SSDataBlock* pBlock) {
+void addTagPseudoColumnData(SReadHandle *pHandle, SExprInfo* pPseudoExpr, int32_t numOfPseudoExpr, SSDataBlock* pBlock) {
   // currently only the tbname pseudo column
-  if (pTableScanInfo->numOfPseudoExpr == 0) {
+  if (numOfPseudoExpr == 0) {
     return;
   }
 
   SMetaReader mr = {0};
-  metaReaderInit(&mr, pTableScanInfo->readHandle.meta, 0);
+  metaReaderInit(&mr, pHandle->meta, 0);
   metaGetTableEntryByUid(&mr, pBlock->info.uid);
 
-  for (int32_t j = 0; j < pTableScanInfo->numOfPseudoExpr; ++j) {
-    SExprInfo* pExpr = &pTableScanInfo->pPseudoExpr[j];
+  for (int32_t j = 0; j < numOfPseudoExpr; ++j) {
+    SExprInfo* pExpr = &pPseudoExpr[j];
 
     int32_t dstSlotId = pExpr->base.resSchema.slotId;
 
@@ -301,7 +309,7 @@ void addTagPseudoColumnData(STableScanInfo* pTableScanInfo, SSDataBlock* pBlock)
 
     // this is to handle the tbname
     if (fmIsScanPseudoColumnFunc(functionId)) {
-      setTbNameColData(pTableScanInfo->readHandle.meta, pBlock, pColInfoData, functionId);
+      setTbNameColData(pHandle->meta, pBlock, pColInfoData, functionId);
     } else {  // these are tags
       const char* p = NULL;
       if (pColInfoData->info.type == TSDB_DATA_TYPE_JSON) {
@@ -380,7 +388,6 @@ static SSDataBlock* doTableScanImpl(SOperatorInfo* pOperator) {
     pOperator->cost.totalCost = pTableScanInfo->readRecorder.elapsedTime;
     return pBlock;
   }
-
   return NULL;
 }
 
@@ -395,9 +402,15 @@ static SSDataBlock* doTableScan(SOperatorInfo* pOperator) {
 
   // do the ascending order traverse in the first place.
   while (pTableScanInfo->scanTimes < pTableScanInfo->scanInfo.numOfAsc) {
-    SSDataBlock* p = doTableScanImpl(pOperator);
-    if (p != NULL) {
-      return p;
+    while (pTableScanInfo->curTWinIdx < pTableScanInfo->cond.numOfTWindows) {
+      SSDataBlock* p = doTableScanImpl(pOperator);
+      if (p != NULL) {
+        return p;
+      }
+      pTableScanInfo->curTWinIdx += 1;
+      if (pTableScanInfo->curTWinIdx < pTableScanInfo->cond.numOfTWindows) {
+        tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond, pTableScanInfo->curTWinIdx);
+      }
     }
 
     pTableScanInfo->scanTimes += 1;
@@ -405,14 +418,14 @@ static SSDataBlock* doTableScan(SOperatorInfo* pOperator) {
     if (pTableScanInfo->scanTimes < pTableScanInfo->scanInfo.numOfAsc) {
       setTaskStatus(pTaskInfo, TASK_NOT_COMPLETED);
       pTableScanInfo->scanFlag = REPEAT_SCAN;
-
-      STimeWindow* pWin = &pTableScanInfo->cond.twindow;
-      qDebug("%s start to repeat ascending order scan data blocks due to query func required, qrange:%" PRId64
-             "-%" PRId64,
-             GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
-
+      qDebug("%s start to repeat ascending order scan data blocks due to query func required", GET_TASKID(pTaskInfo));
+      for (int32_t i = 0; i < pTableScanInfo->cond.numOfTWindows; ++i) {
+        STimeWindow* pWin = &pTableScanInfo->cond.twindows[i];
+        qDebug("%s\t qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
+      }
       // do prepare for the next round table scan operation
-      tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond);
+      tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond, 0);
+      pTableScanInfo->curTWinIdx = 0;
     }
   }
 
@@ -420,31 +433,40 @@ static SSDataBlock* doTableScan(SOperatorInfo* pOperator) {
   if (pTableScanInfo->scanTimes < total) {
     if (pTableScanInfo->cond.order == TSDB_ORDER_ASC) {
       prepareForDescendingScan(pTableScanInfo, pTableScanInfo->pCtx, pTableScanInfo->numOfOutput);
-      tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond);
+      tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond, 0);
+      pTableScanInfo->curTWinIdx = 0;
     }
 
-    STimeWindow* pWin = &pTableScanInfo->cond.twindow;
-    qDebug("%s start to descending order scan data blocks due to query func required, qrange:%" PRId64 "-%" PRId64,
-           GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
-
+    qDebug("%s start to descending order scan data blocks due to query func required", GET_TASKID(pTaskInfo));
+    for (int32_t i = 0; i < pTableScanInfo->cond.numOfTWindows; ++i) {
+      STimeWindow* pWin = &pTableScanInfo->cond.twindows[i];
+      qDebug("%s\t qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
+    }
     while (pTableScanInfo->scanTimes < total) {
-      SSDataBlock* p = doTableScanImpl(pOperator);
-      if (p != NULL) {
-        return p;
+      while (pTableScanInfo->curTWinIdx < pTableScanInfo->cond.numOfTWindows) {
+        SSDataBlock* p = doTableScanImpl(pOperator);
+        if (p != NULL) {
+          return p;
+        }
+        pTableScanInfo->curTWinIdx += 1;
+        if (pTableScanInfo->curTWinIdx < pTableScanInfo->cond.numOfTWindows) {
+          tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond, pTableScanInfo->curTWinIdx);
+        }
       }
 
       pTableScanInfo->scanTimes += 1;
 
-      if (pTableScanInfo->scanTimes < pTableScanInfo->scanInfo.numOfAsc) {
+      if (pTableScanInfo->scanTimes < total) {
         setTaskStatus(pTaskInfo, TASK_NOT_COMPLETED);
         pTableScanInfo->scanFlag = REPEAT_SCAN;
 
-        qDebug("%s start to repeat descending order scan data blocks due to query func required, qrange:%" PRId64
-               "-%" PRId64,
-               GET_TASKID(pTaskInfo), pTaskInfo->window.skey, pTaskInfo->window.ekey);
-
-        // do prepare for the next round table scan operation
-        tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond);
+        qDebug("%s start to repeat descending order scan data blocks due to query func required", GET_TASKID(pTaskInfo));
+        for (int32_t i = 0; i < pTableScanInfo->cond.numOfTWindows; ++i) {
+          STimeWindow* pWin = &pTableScanInfo->cond.twindows[i];
+          qDebug("%s\t qrange:%" PRId64 "-%" PRId64, GET_TASKID(pTaskInfo), pWin->skey, pWin->ekey);
+        }
+        tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond, 0);
+        pTableScanInfo->curTWinIdx = 0;
       }
     }
   }
@@ -524,6 +546,7 @@ SOperatorInfo* createTableScanOperatorInfo(STableScanPhysiNode* pTableScanNode,
   pInfo->dataReader = pDataReader;
   pInfo->scanFlag = MAIN_SCAN;
   pInfo->pColMatchInfo = pColList;
+  pInfo->curTWinIdx = 0;
 
   pOperator->name = "TableScanOperator";  // for debug purpose
   pOperator->operatorType = QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN;
@@ -678,8 +701,9 @@ static bool prepareDataScan(SStreamBlockScanInfo* pInfo) {
                                                         binarySearchForKey, NULL, TSDB_ORDER_ASC);
     }
     STableScanInfo* pTableScanInfo = pInfo->pOperatorDumy->info;
-    pTableScanInfo->cond.twindow = win;
-    tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond);
+    pTableScanInfo->cond.twindows[0] = win;
+    pTableScanInfo->curTWinIdx = 0;
+    tsdbResetReadHandle(pTableScanInfo->dataReader, &pTableScanInfo->cond, 0);
     pTableScanInfo->scanTimes = 0;
     return true;
   } else {
@@ -731,91 +755,6 @@ static SSDataBlock* getUpdateDataBlock(SStreamBlockScanInfo* pInfo, bool inverti
   return NULL;
 }
 
-void static setSupKeyBuf(SCatchSupporter* pSup, int64_t groupId, int64_t childId, TSKEY ts) {
-  int64_t* pKey = (int64_t*)pSup->pKeyBuf;
-  pKey[0] = groupId;
-  pKey[1] = childId;
-  pKey[2] = ts;
-}
-
-static int32_t catchWidonwInfo(SSDataBlock* pDataBlock, SCatchSupporter* pSup, int32_t pageId, int32_t tsIndex,
-                               int64_t childId) {
-  SColumnInfoData* pColDataInfo = taosArrayGet(pDataBlock->pDataBlock, tsIndex);
-  TSKEY*           tsCols = (int64_t*)pColDataInfo->pData;
-  for (int32_t i = 0; i < pDataBlock->info.rows; i++) {
-    setSupKeyBuf(pSup, pDataBlock->info.groupId, childId, tsCols[i]);
-    SWindowPosition* p1 = (SWindowPosition*)taosHashGet(pSup->pWindowHashTable, pSup->pKeyBuf, pSup->keySize);
-    if (p1 == NULL) {
-      SWindowPosition pos = {.pageId = pageId, .rowId = i};
-      int32_t code = taosHashPut(pSup->pWindowHashTable, pSup->pKeyBuf, pSup->keySize, &pos, sizeof(SWindowPosition));
-      if (code != TSDB_CODE_SUCCESS) {
-        return code;
-      }
-    } else {
-      p1->pageId = pageId;
-      p1->rowId = i;
-    }
-  }
-  return TSDB_CODE_SUCCESS;
-}
-
-static int32_t catchDatablock(SSDataBlock* pDataBlock, SCatchSupporter* pSup, int32_t tsIndex, int64_t childId) {
-  int32_t start = 0;
-  int32_t stop = 0;
-  int32_t pageSize = getBufPageSize(pSup->pDataBuf);
-  while (start < pDataBlock->info.rows) {
-    blockDataSplitRows(pDataBlock, pDataBlock->info.hasVarCol, start, &stop, pageSize);
-    SSDataBlock* pDB = blockDataExtractBlock(pDataBlock, start, stop - start + 1);
-    if (pDB == NULL) {
-      return terrno;
-    }
-    int32_t pageId = -1;
-    void*   pPage = getNewBufPage(pSup->pDataBuf, pDataBlock->info.groupId, &pageId);
-    if (pPage == NULL) {
-      blockDataDestroy(pDB);
-      return terrno;
-    }
-    int32_t size = blockDataGetSize(pDB) + sizeof(int32_t) + pDB->info.numOfCols * sizeof(int32_t);
-    assert(size <= pageSize);
-    blockDataToBuf(pPage, pDB);
-    setBufPageDirty(pPage, true);
-    releaseBufPage(pSup->pDataBuf, pPage);
-    blockDataDestroy(pDB);
-    start = stop + 1;
-    int32_t code = catchWidonwInfo(pDB, pSup, pageId, tsIndex, childId);
-    if (code != TSDB_CODE_SUCCESS) {
-      return code;
-    }
-  }
-  return TSDB_CODE_SUCCESS;
-}
-
-static SSDataBlock* getDataFromCatch(SStreamBlockScanInfo* pInfo) {
-  SSDataBlock* pBlock = pInfo->pUpdateRes;
-  if (pInfo->updateResIndex < pBlock->info.rows) {
-    blockDataCleanup(pInfo->pRes);
-    SCatchSupporter* pCSup = &pInfo->childAggSup;
-    SColumnInfoData* pColDataInfo = taosArrayGet(pBlock->pDataBlock, 0);
-    TSKEY*           tsCols = (TSKEY*)pColDataInfo->pData;
-    int32_t          size = taosArrayGetSize(pInfo->childIds);
-    for (int32_t i = 0; i < size; i++) {
-      int64_t id = *(int64_t*)taosArrayGet(pInfo->childIds, i);
-      setSupKeyBuf(pCSup, pBlock->info.groupId, id, tsCols[pInfo->updateResIndex]);
-      SWindowPosition* pos = (SWindowPosition*)taosHashGet(pCSup->pWindowHashTable, pCSup->pKeyBuf, pCSup->keySize);
-      void*            buf = getBufPage(pCSup->pDataBuf, pos->pageId);
-      SSDataBlock*     pDB = createOneDataBlock(pInfo->pRes, false);
-      blockDataFromBuf(pDB, buf);
-      SSDataBlock* pSub = blockDataExtractBlock(pDB, pos->rowId, 1);
-      blockDataMerge(pInfo->pRes, pSub);
-      blockDataDestroy(pDB);
-      blockDataDestroy(pSub);
-    }
-    pInfo->updateResIndex++;
-    return pInfo->pRes;
-  }
-  return NULL;
-}
-
 static SSDataBlock* doStreamBlockScan(SOperatorInfo* pOperator) {
   // NOTE: this operator does never check if current status is done or not
   SExecTaskInfo*        pTaskInfo = pOperator->pTaskInfo;
@@ -886,8 +825,7 @@ static SSDataBlock* doStreamBlockScan(SOperatorInfo* pOperator) {
         pInfo->pRes->info.groupId = groupId;
       }
 
-      int32_t numOfCols = pInfo->pRes->info.numOfCols;
-      for (int32_t i = 0; i < numOfCols; ++i) {
+      for (int32_t i = 0; i < taosArrayGetSize(pInfo->pColMatchInfo); ++i) {
         SColMatchInfo* pColMatchInfo = taosArrayGet(pInfo->pColMatchInfo, i);
         if (!pColMatchInfo->output) {
           continue;
@@ -917,10 +855,16 @@ static SSDataBlock* doStreamBlockScan(SOperatorInfo* pOperator) {
         pTaskInfo->code = terrno;
         return NULL;
       }
+
       rows = pBlockInfo->rows;
+
+      // currently only the tbname pseudo column
+      if (pInfo->numOfPseudoExpr > 0) {
+        addTagPseudoColumnData(&pInfo->readHandle, pInfo->pPseudoExpr, pInfo->numOfPseudoExpr, pInfo->pRes);
+      }
+
       doFilter(pInfo->pCondition, pInfo->pRes, NULL);
       blockDataUpdateTsWindow(pInfo->pRes, 0);
-
       break;
     }
 
@@ -948,10 +892,9 @@ static SSDataBlock* doStreamBlockScan(SOperatorInfo* pOperator) {
   }
 }
 
-SOperatorInfo* createStreamScanOperatorInfo(void* streamReadHandle, void* pDataReader, SReadHandle* pHandle,
-                                            uint64_t uid, SSDataBlock* pResBlock, SArray* pColList,
-                                            SArray* pTableIdList, SExecTaskInfo* pTaskInfo, SNode* pCondition,
-                                            SOperatorInfo* pOperatorDumy) {
+SOperatorInfo* createStreamScanOperatorInfo(void* pDataReader, SReadHandle* pHandle,
+    SArray* pTableIdList, STableScanPhysiNode* pTableScanNode, SExecTaskInfo* pTaskInfo,
+    STimeWindowAggSupp* pTwSup, int16_t tsColId) {
   SStreamBlockScanInfo* pInfo = taosMemoryCalloc(1, sizeof(SStreamBlockScanInfo));
   SOperatorInfo*        pOperator = taosMemoryCalloc(1, sizeof(SOperatorInfo));
   if (pInfo == NULL || pOperator == NULL) {
@@ -959,22 +902,28 @@ SOperatorInfo* createStreamScanOperatorInfo(void* streamReadHandle, void* pDataR
     goto _error;
   }
 
-  STableScanInfo* pSTInfo = (STableScanInfo*)pOperatorDumy->info;
+  SScanPhysiNode* pScanPhyNode = &pTableScanNode->scan;
+
+  SDataBlockDescNode* pDescNode = pScanPhyNode->node.pOutputDataBlockDesc;
+  SOperatorInfo*      pTableScanDummy = createTableScanOperatorInfo(pTableScanNode, pDataReader, pHandle, pTaskInfo);
+
+  STableScanInfo* pSTInfo = (STableScanInfo*)pTableScanDummy->info;
 
-  int32_t numOfOutput = taosArrayGetSize(pColList);
+  int32_t numOfCols = 0;
+  pInfo->pColMatchInfo = extractColMatchInfo(pScanPhyNode->pScanCols, pDescNode, &numOfCols, pTaskInfo, COL_MATCH_FROM_COL_ID);
 
-  SArray* pColIds = taosArrayInit(4, sizeof(int16_t));
+  int32_t numOfOutput = taosArrayGetSize(pInfo->pColMatchInfo);
+  SArray* pColIds = taosArrayInit(numOfOutput, sizeof(int16_t));
   for (int32_t i = 0; i < numOfOutput; ++i) {
-    SColMatchInfo* id = taosArrayGet(pColList, i);
-    int16_t        colId = id->colId;
+    SColMatchInfo* id = taosArrayGet(pInfo->pColMatchInfo, i);
+
+    int16_t colId = id->colId;
     taosArrayPush(pColIds, &colId);
   }
 
-  pInfo->pColMatchInfo = pColList;
-
   // set the extract column id to streamHandle
-  tqReadHandleSetColIdList((STqReadHandle*)streamReadHandle, pColIds);
-  int32_t code = tqReadHandleSetTbUidList(streamReadHandle, pTableIdList);
+  tqReadHandleSetColIdList((STqReadHandle*)pHandle->reader, pColIds);
+  int32_t code = tqReadHandleSetTbUidList(pHandle->reader, pTableIdList);
   if (code != 0) {
     goto _error;
   }
@@ -990,37 +939,39 @@ SOperatorInfo* createStreamScanOperatorInfo(void* streamReadHandle, void* pDataR
     goto _error;
   }
 
-  pInfo->primaryTsIndex = 0;  // TODO(liuyao) get it from physical plan
+  pInfo->primaryTsIndex = tsColId;
   if (pSTInfo->interval.interval > 0) {
-    pInfo->pUpdateInfo = updateInfoInitP(&pSTInfo->interval, 10000);  // TODO(liuyao) get watermark from physical plan
+    pInfo->pUpdateInfo = updateInfoInitP(&pSTInfo->interval, pTwSup->waterMark);
   } else {
     pInfo->pUpdateInfo = NULL;
   }
 
-  pInfo->readHandle = *pHandle;
-  pInfo->tableUid = uid;
-  pInfo->streamBlockReader = streamReadHandle;
-  pInfo->pRes = pResBlock;
-  pInfo->pCondition = pCondition;
-  pInfo->pDataReader = pDataReader;
-  pInfo->scanMode = STREAM_SCAN_FROM_READERHANDLE;
-  pInfo->pOperatorDumy = pOperatorDumy;
-  pInfo->interval = pSTInfo->interval;
-  pInfo->sessionSup = (SessionWindowSupporter){.pStreamAggSup = NULL, .gap = -1};
-
-  initCacheSupporter(&pInfo->childAggSup, 1024, "StreamFinalInterval",
-                     "/tmp/");  // TODO(liuyao) get row size from phy plan
+  // create the pseduo columns info
+  if (pTableScanNode->scan.pScanPseudoCols != NULL) {
+    pInfo->pPseudoExpr = createExprInfo(pTableScanNode->scan.pScanPseudoCols, NULL, &pInfo->numOfPseudoExpr);
+  }
 
-  pOperator->name = "StreamBlockScanOperator";
+  pInfo->readHandle        = *pHandle;
+  pInfo->tableUid          = pScanPhyNode->uid;
+  pInfo->streamBlockReader = pHandle->reader;
+  pInfo->pRes              = createResDataBlock(pDescNode);
+  pInfo->pCondition        = pScanPhyNode->node.pConditions;
+  pInfo->pDataReader       = pDataReader;
+  pInfo->scanMode          = STREAM_SCAN_FROM_READERHANDLE;
+  pInfo->pOperatorDumy     = pTableScanDummy;
+  pInfo->interval          = pSTInfo->interval;
+  pInfo->sessionSup        = (SessionWindowSupporter){.pStreamAggSup = NULL, .gap = -1};
+
+  pOperator->name         = "StreamBlockScanOperator";
   pOperator->operatorType = QUERY_NODE_PHYSICAL_PLAN_STREAM_SCAN;
-  pOperator->blocking = false;
-  pOperator->status = OP_NOT_OPENED;
-  pOperator->info = pInfo;
-  pOperator->numOfExprs = pResBlock->info.numOfCols;
-  pOperator->pTaskInfo = pTaskInfo;
-
-  pOperator->fpSet =
-      createOperatorFpSet(operatorDummyOpenFn, doStreamBlockScan, NULL, NULL, operatorDummyCloseFn, NULL, NULL, NULL);
+  pOperator->blocking     = false;
+  pOperator->status       = OP_NOT_OPENED;
+  pOperator->info         = pInfo;
+  pOperator->numOfExprs   = pInfo->pRes->info.numOfCols;
+  pOperator->pTaskInfo    = pTaskInfo;
+
+  pOperator->fpSet = createOperatorFpSet(operatorDummyOpenFn, doStreamBlockScan, NULL,
+      NULL, operatorDummyCloseFn, NULL, NULL, NULL);
 
   return pOperator;
 

source/libs/function/src/builtins.c

@@ -156,6 +156,14 @@ static int32_t translatePercentile(SFunctionNode* pFunc, char* pErrBuf, int32_t
     return invaildFuncParaNumErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
+  //param0
+  SNode* pParamNode0 = nodesListGetNode(pFunc->pParameterList, 0);
+  if (nodeType(pParamNode0) != QUERY_NODE_COLUMN) {
+    return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_ERROR,
+                           "The first parameter of PERCENTILE function can only be column");
+  }
+
+  //param1
   SValueNode* pValue = (SValueNode*)nodesListGetNode(pFunc->pParameterList, 1);
 
   if (pValue->datum.i < 0 || pValue->datum.i > 100) {
@@ -170,6 +178,7 @@ static int32_t translatePercentile(SFunctionNode* pFunc, char* pErrBuf, int32_t
     return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
+  //set result type
   pFunc->node.resType = (SDataType){.bytes = tDataTypes[TSDB_DATA_TYPE_DOUBLE].bytes, .type = TSDB_DATA_TYPE_DOUBLE};
   return TSDB_CODE_SUCCESS;
 }
@@ -188,30 +197,47 @@ static int32_t translateApercentile(SFunctionNode* pFunc, char* pErrBuf, int32_t
     return invaildFuncParaNumErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
-  uint8_t para1Type = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 0))->resType.type;
-  uint8_t para2Type = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 1))->resType.type;
-  if (!IS_NUMERIC_TYPE(para1Type) || !IS_INTEGER_TYPE(para2Type)) {
-    return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
+  //param0
+  SNode* pParamNode0 = nodesListGetNode(pFunc->pParameterList, 0);
+  if (nodeType(pParamNode0) != QUERY_NODE_COLUMN) {
+    return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_ERROR,
+                           "The first parameter of APERCENTILE function can only be column");
   }
 
-  SNode* pParamNode = nodesListGetNode(pFunc->pParameterList, 1);
-  if (nodeType(pParamNode) != QUERY_NODE_VALUE) {
+  //param1
+  SNode* pParamNode1 = nodesListGetNode(pFunc->pParameterList, 1);
+  if (nodeType(pParamNode1) != QUERY_NODE_VALUE) {
     return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
-  SValueNode* pValue = (SValueNode*)pParamNode;
+  SValueNode* pValue = (SValueNode*)pParamNode1;
   if (pValue->datum.i < 0 || pValue->datum.i > 100) {
     return invaildFuncParaValueErrMsg(pErrBuf, len, pFunc->functionName);
   }
 
   pValue->notReserved = true;
 
+  uint8_t para1Type = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 0))->resType.type;
+  uint8_t para2Type = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 1))->resType.type;
+  if (!IS_NUMERIC_TYPE(para1Type) || !IS_INTEGER_TYPE(para2Type)) {
+    return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
+  }
+
+  //param2
   if (3 == numOfParams) {
-    SNode* pPara3 = nodesListGetNode(pFunc->pParameterList, 2);
-    if (QUERY_NODE_VALUE != nodeType(pPara3) || !validAperventileAlgo((SValueNode*)pPara3)) {
+    uint8_t para3Type = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 2))->resType.type;
+    if (!IS_VAR_DATA_TYPE(para3Type)) {
+      return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
+    }
+
+    SNode* pParamNode2 = nodesListGetNode(pFunc->pParameterList, 2);
+    if (QUERY_NODE_VALUE != nodeType(pParamNode2) || !validAperventileAlgo((SValueNode*)pParamNode2)) {
       return buildFuncErrMsg(pErrBuf, len, TSDB_CODE_FUNC_FUNTION_ERROR,
                              "Third parameter algorithm of apercentile must be 'default' or 't-digest'");
     }
+
+    pValue = (SValueNode*)pParamNode2;
+    pValue->notReserved = true;
   }
 
   pFunc->node.resType = (SDataType){.bytes = tDataTypes[TSDB_DATA_TYPE_DOUBLE].bytes, .type = TSDB_DATA_TYPE_DOUBLE};
@@ -700,6 +726,11 @@ static int32_t translateDiff(SFunctionNode* pFunc, char* pErrBuf, int32_t len) {
 
   //param1
   if (numOfParams == 2) {
+    uint8_t paraType = ((SExprNode*)nodesListGetNode(pFunc->pParameterList, 1))->resType.type;
+    if (!IS_INTEGER_TYPE(paraType)) {
+      return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);
+    }
+
     SNode* pParamNode1 = nodesListGetNode(pFunc->pParameterList, 1);
     if (QUERY_NODE_VALUE != nodeType(pParamNode1)) {
       return invaildFuncParaTypeErrMsg(pErrBuf, len, pFunc->functionName);

source/libs/function/src/builtinsimpl.c

@@ -1965,7 +1965,7 @@ bool apercentileFunctionSetup(SqlFunctionCtx* pCtx, SResultRowEntryInfo* pResult
   if (pCtx->numOfParams == 2) {
     pInfo->algo = APERCT_ALGO_DEFAULT;
   } else if (pCtx->numOfParams == 3) {
-    pInfo->algo = getApercentileAlgo(pCtx->param[2].param.pz);
+    pInfo->algo = getApercentileAlgo(varDataVal(pCtx->param[2].param.pz));
     if (pInfo->algo == APERCT_ALGO_UNKNOWN) {
       return false;
     }

source/libs/index/inc/indexCache.h

@@ -74,7 +74,7 @@ void     indexCacheIteratorDestroy(Iterate* iiter);
 int indexCachePut(void* cache, SIndexTerm* term, uint64_t uid);
 
 // int indexCacheGet(void *cache, uint64_t *rst);
-int indexCacheSearch(void* cache, SIndexTermQuery* query, SIdxTempResult* tr, STermValueType* s);
+int indexCacheSearch(void* cache, SIndexTermQuery* query, SIdxTRslt* tr, STermValueType* s);
 
 void indexCacheRef(IndexCache* cache);
 void indexCacheUnRef(IndexCache* cache);

source/libs/index/inc/indexTfile.h

@@ -105,7 +105,7 @@ TFileReader* tfileGetReaderByCol(IndexTFile* tf, uint64_t suid, char* colName);
 TFileReader* tfileReaderOpen(char* path, uint64_t suid, int64_t version, const char* colName);
 TFileReader* tfileReaderCreate(WriterCtx* ctx);
 void         tfileReaderDestroy(TFileReader* reader);
-int          tfileReaderSearch(TFileReader* reader, SIndexTermQuery* query, SIdxTempResult* tr);
+int          tfileReaderSearch(TFileReader* reader, SIndexTermQuery* query, SIdxTRslt* tr);
 void         tfileReaderRef(TFileReader* reader);
 void         tfileReaderUnRef(TFileReader* reader);
 
@@ -120,7 +120,7 @@ int          tfileWriterFinish(TFileWriter* tw);
 IndexTFile* indexTFileCreate(const char* path);
 void        indexTFileDestroy(IndexTFile* tfile);
 int         indexTFilePut(void* tfile, SIndexTerm* term, uint64_t uid);
-int         indexTFileSearch(void* tfile, SIndexTermQuery* query, SIdxTempResult* tr);
+int         indexTFileSearch(void* tfile, SIndexTermQuery* query, SIdxTRslt* tr);
 
 Iterate* tfileIteratorCreate(TFileReader* reader);
 void     tfileIteratorDestroy(Iterate* iterator);

source/libs/index/inc/indexUtil.h

@@ -66,7 +66,7 @@ extern "C" {
  *        [1, 4, 5]
  * output:[4, 5]
  */
-void iIntersection(SArray *interResults, SArray *finalResult);
+void iIntersection(SArray *in, SArray *out);
 
 /* multi sorted result union
  * input: [1, 2, 4, 5]
@@ -74,7 +74,7 @@ void iIntersection(SArray *interResults, SArray *finalResult);
  *        [1, 4, 5]
  * output:[1, 2, 3, 4, 5]
  */
-void iUnion(SArray *interResults, SArray *finalResult);
+void iUnion(SArray *in, SArray *out);
 
 /* see example
  * total:   [1, 2, 4, 5, 7, 8]
@@ -92,19 +92,24 @@ typedef struct {
   uint64_t data;
 } SIdxVerdata;
 
+/*
+ * index temp result
+ *
+ */
 typedef struct {
   SArray *total;
-  SArray *added;
-  SArray *deled;
-} SIdxTempResult;
+  SArray *add;
+  SArray *del;
+} SIdxTRslt;
+
+SIdxTRslt *idxTRsltCreate();
 
-SIdxTempResult *sIdxTempResultCreate();
+void idxTRsltClear(SIdxTRslt *tr);
 
-void sIdxTempResultClear(SIdxTempResult *tr);
+void idxTRsltDestroy(SIdxTRslt *tr);
 
-void sIdxTempResultDestroy(SIdxTempResult *tr);
+void idxTRsltMergeTo(SIdxTRslt *tr, SArray *out);
 
-void sIdxTempResultMergeTo(SArray *result, SIdxTempResult *tr);
 #ifdef __cplusplus
 }
 #endif

source/libs/index/src/index.c

@@ -29,7 +29,7 @@
 #include "lucene++/Lucene_c.h"
 #endif
 
-#define INDEX_NUM_OF_THREADS 1
+#define INDEX_NUM_OF_THREADS 5
 #define INDEX_QUEUE_SIZE     200
 
 #define INDEX_DATA_BOOL_NULL      0x02
@@ -85,7 +85,7 @@ static int  indexMergeFinalResults(SArray* interResults, EIndexOperatorType oTyp
 static int indexGenTFile(SIndex* index, IndexCache* cache, SArray* batch);
 
 // merge cache and tfile by opera type
-static void indexMergeCacheAndTFile(SArray* result, IterateValue* icache, IterateValue* iTfv, SIdxTempResult* helper);
+static void indexMergeCacheAndTFile(SArray* result, IterateValue* icache, IterateValue* iTfv, SIdxTRslt* helper);
 
 // static int32_t indexSerialTermKey(SIndexTerm* itm, char* buf);
 // int32_t        indexSerialKey(ICacheKey* key, char* buf);
@@ -201,6 +201,7 @@ int indexPut(SIndex* index, SIndexMultiTerm* fVals, uint64_t uid) {
     char      buf[128] = {0};
     ICacheKey key = {.suid = p->suid, .colName = p->colName, .nColName = strlen(p->colName), .colType = p->colType};
     int32_t   sz = indexSerialCacheKey(&key, buf);
+    indexDebug("suid: %" PRIu64 ", colName: %s, colType: %d", key.suid, key.colName, key.colType);
 
     IndexCache** cache = taosHashGet(index->colObj, buf, sz);
     assert(*cache != NULL);
@@ -328,6 +329,7 @@ static int indexTermSearch(SIndex* sIdx, SIndexTermQuery* query, SArray** result
   char      buf[128] = {0};
   ICacheKey key = {
       .suid = term->suid, .colName = term->colName, .nColName = strlen(term->colName), .colType = term->colType};
+  indexDebug("suid: %" PRIu64 ", colName: %s, colType: %d", key.suid, key.colName, key.colType);
   int32_t sz = indexSerialCacheKey(&key, buf);
 
   taosThreadMutexLock(&sIdx->mtx);
@@ -341,7 +343,7 @@ static int indexTermSearch(SIndex* sIdx, SIndexTermQuery* query, SArray** result
 
   int64_t st = taosGetTimestampUs();
 
-  SIdxTempResult* tr = sIdxTempResultCreate();
+  SIdxTRslt* tr = idxTRsltCreate();
   if (0 == indexCacheSearch(cache, query, tr, &s)) {
     if (s == kTypeDeletion) {
       indexInfo("col: %s already drop by", term->colName);
@@ -363,12 +365,12 @@ static int indexTermSearch(SIndex* sIdx, SIndexTermQuery* query, SArray** result
   int64_t cost = taosGetTimestampUs() - st;
   indexInfo("search cost: %" PRIu64 "us", cost);
 
-  sIdxTempResultMergeTo(*result, tr);
+  idxTRsltMergeTo(tr, *result);
 
-  sIdxTempResultDestroy(tr);
+  idxTRsltDestroy(tr);
   return 0;
 END:
-  sIdxTempResultDestroy(tr);
+  idxTRsltDestroy(tr);
   return -1;
 }
 static void indexInterResultsDestroy(SArray* results) {
@@ -404,18 +406,18 @@ static int indexMergeFinalResults(SArray* interResults, EIndexOperatorType oType
   return 0;
 }
 
-static void indexMayMergeTempToFinalResult(SArray* result, TFileValue* tfv, SIdxTempResult* tr) {
+static void indexMayMergeTempToFinalResult(SArray* result, TFileValue* tfv, SIdxTRslt* tr) {
   int32_t sz = taosArrayGetSize(result);
   if (sz > 0) {
     TFileValue* lv = taosArrayGetP(result, sz - 1);
     if (tfv != NULL && strcmp(lv->colVal, tfv->colVal) != 0) {
-      sIdxTempResultMergeTo(lv->tableId, tr);
-      sIdxTempResultClear(tr);
+      idxTRsltMergeTo(tr, lv->tableId);
+      idxTRsltClear(tr);
 
       taosArrayPush(result, &tfv);
     } else if (tfv == NULL) {
       // handle last iterator
-      sIdxTempResultMergeTo(lv->tableId, tr);
+      idxTRsltMergeTo(tr, lv->tableId);
     } else {
       // temp result saved in help
       tfileValueDestroy(tfv);
@@ -424,7 +426,7 @@ static void indexMayMergeTempToFinalResult(SArray* result, TFileValue* tfv, SIdx
     taosArrayPush(result, &tfv);
   }
 }
-static void indexMergeCacheAndTFile(SArray* result, IterateValue* cv, IterateValue* tv, SIdxTempResult* tr) {
+static void indexMergeCacheAndTFile(SArray* result, IterateValue* cv, IterateValue* tv, SIdxTRslt* tr) {
   char*       colVal = (cv != NULL) ? cv->colVal : tv->colVal;
   TFileValue* tfv = tfileValueCreate(colVal);
 
@@ -434,9 +436,9 @@ static void indexMergeCacheAndTFile(SArray* result, IterateValue* cv, IterateVal
     uint64_t id = *(uint64_t*)taosArrayGet(cv->val, 0);
     uint32_t ver = cv->ver;
     if (cv->type == ADD_VALUE) {
-      INDEX_MERGE_ADD_DEL(tr->deled, tr->added, id)
+      INDEX_MERGE_ADD_DEL(tr->del, tr->add, id)
     } else if (cv->type == DEL_VALUE) {
-      INDEX_MERGE_ADD_DEL(tr->added, tr->deled, id)
+      INDEX_MERGE_ADD_DEL(tr->add, tr->del, id)
     }
   }
   if (tv != NULL) {
@@ -489,7 +491,7 @@ int indexFlushCacheToTFile(SIndex* sIdx, void* cache) {
   bool cn = cacheIter ? cacheIter->next(cacheIter) : false;
   bool tn = tfileIter ? tfileIter->next(tfileIter) : false;
 
-  SIdxTempResult* tr = sIdxTempResultCreate();
+  SIdxTRslt* tr = idxTRsltCreate();
   while (cn == true || tn == true) {
     IterateValue* cv = (cn == true) ? cacheIter->getValue(cacheIter) : NULL;
     IterateValue* tv = (tn == true) ? tfileIter->getValue(tfileIter) : NULL;
@@ -515,7 +517,7 @@ int indexFlushCacheToTFile(SIndex* sIdx, void* cache) {
     }
   }
   indexMayMergeTempToFinalResult(result, NULL, tr);
-  sIdxTempResultDestroy(tr);
+  idxTRsltDestroy(tr);
 
   int ret = indexGenTFile(sIdx, pCache, result);
   indexDestroyFinalResult(result);

source/libs/index/test/indexTests.cc

@@ -411,12 +411,12 @@ class TFileObj {
       //
       //
     }
-    SIdxTempResult* tr = sIdxTempResultCreate();
+    SIdxTRslt* tr = idxTRsltCreate();
 
     int ret = tfileReaderSearch(reader_, query, tr);
 
-    sIdxTempResultMergeTo(result, tr);
-    sIdxTempResultDestroy(tr);
+    idxTRsltMergeTo(tr, result);
+    idxTRsltDestroy(tr);
     return ret;
   }
   ~TFileObj() {
@@ -531,11 +531,11 @@ class CacheObj {
     indexCacheDebug(cache);
   }
   int Get(SIndexTermQuery* query, int16_t colId, int32_t version, SArray* result, STermValueType* s) {
-    SIdxTempResult* tr = sIdxTempResultCreate();
+    SIdxTRslt* tr = idxTRsltCreate();
 
     int ret = indexCacheSearch(cache, query, tr, s);
-    sIdxTempResultMergeTo(result, tr);
-    sIdxTempResultDestroy(tr);
+    idxTRsltMergeTo(tr, result);
+    idxTRsltDestroy(tr);
 
     if (ret != 0) {
       std::cout << "failed to get from cache:" << ret << std::endl;

source/libs/nodes/src/nodesCodeFuncs.c

@@ -1130,6 +1130,9 @@ static const char* jkTableScanPhysiPlanOffset = "Offset";
 static const char* jkTableScanPhysiPlanSliding = "Sliding";
 static const char* jkTableScanPhysiPlanIntervalUnit = "intervalUnit";
 static const char* jkTableScanPhysiPlanSlidingUnit = "slidingUnit";
+static const char* jkTableScanPhysiPlanTriggerType = "triggerType";
+static const char* jkTableScanPhysiPlanWatermark = "watermark";
+static const char* jkTableScanPhysiPlanTsColId = "tsColId";
 
 static int32_t physiTableScanNodeToJson(const void* pObj, SJson* pJson) {
   const STableScanPhysiNode* pNode = (const STableScanPhysiNode*)pObj;
@@ -1171,6 +1174,15 @@ static int32_t physiTableScanNodeToJson(const void* pObj, SJson* pJson) {
   if (TSDB_CODE_SUCCESS == code) {
     code = tjsonAddIntegerToObject(pJson, jkTableScanPhysiPlanSlidingUnit, pNode->slidingUnit);
   }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddIntegerToObject(pJson, jkTableScanPhysiPlanTriggerType, pNode->triggerType);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddIntegerToObject(pJson, jkTableScanPhysiPlanWatermark, pNode->watermark);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    code = tjsonAddIntegerToObject(pJson, jkTableScanPhysiPlanTsColId, pNode->tsColId);
+  }
 
   return code;
 }
@@ -1221,6 +1233,15 @@ static int32_t jsonToPhysiTableScanNode(const SJson* pJson, void* pObj) {
     tjsonGetNumberValue(pJson, jkTableScanPhysiPlanSlidingUnit, pNode->slidingUnit, code);
     ;
   }
+  if (TSDB_CODE_SUCCESS == code) {
+    tjsonGetNumberValue(pJson, jkTableScanPhysiPlanTriggerType, pNode->triggerType, code);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    tjsonGetNumberValue(pJson, jkTableScanPhysiPlanWatermark, pNode->watermark, code);
+  }
+  if (TSDB_CODE_SUCCESS == code) {
+    tjsonGetNumberValue(pJson, jkTableScanPhysiPlanTsColId, pNode->tsColId, code);
+  }
 
   return code;
 }

source/libs/planner/src/planOptimizer.c

@@ -223,6 +223,9 @@ static void setScanWindowInfo(SScanLogicNode* pScan) {
     pScan->sliding = ((SWindowLogicNode*)pScan->node.pParent)->sliding;
     pScan->intervalUnit = ((SWindowLogicNode*)pScan->node.pParent)->intervalUnit;
     pScan->slidingUnit = ((SWindowLogicNode*)pScan->node.pParent)->slidingUnit;
+    pScan->triggerType = ((SWindowLogicNode*)pScan->node.pParent)->triggerType;
+    pScan->watermark = ((SWindowLogicNode*)pScan->node.pParent)->watermark;
+    pScan->tsColId = ((SColumnNode*)((SWindowLogicNode*)pScan->node.pParent)->pTspk)->colId;
   }
 }
 

tests/script/tsim/stable/alter_count.sim

@@ -141,6 +141,8 @@ sql connect
 
 sql select count(a), count(b), count(c), count(d), count(e), count(f), count(g), count(h) from d1.tb;
 sql select count(a), count(b), count(c), count(d), count(e), count(f), count(g), count(h) from d1.tb;
+
+sql use d1
 sql select count(a), count(b), count(c), count(d), count(e), count(f), count(g), count(h) from tb
 if $data00 != 24 then
   return -1