This topic describes the engine release notes for AnalyticDB for PostgreSQL in elastic storage mode in 2022 and provides links to the relevant references.

You can update your instances to the latest minor engine version in the AnalyticDB for PostgreSQL console. For more information about how to update the minor engine version, see Update the minor engine version.

July 28, 2022 (V6.3.10.1)

Category Feature Description References
New feature Diagnostics and optimization The following pages are added for the diagnostics and optimization feature in the console:
  • Overview: allows you to view the connection status, cluster status, and storage status of your instance.
  • Key Metrics: allows you to view metrics about data bloat, data skew, and index usage statistics.
  • Query Analysis: allows you to view the execution information and status of SQL queries and use execution plan details to analyze the query execution process.
Real-time materialized view The following features are supported:
  • OUTER JOIN for real-time materialized views.
  • Nesting for real-time materialized views.
  • Query rewrite for materialized views.
Real-time materialized views
Optimized feature Real-time materialized view The write performance of real-time materialized views is optimized. Real-time materialized views
Fixed issue The issue that may cause data write failures when a real-time materialized view and its base table are used by different users is fixed.
The issue that may cause errors if the query rewrite for a materialized view involves window functions is fixed.

April 15, 2022 (V6.3.9.0)

Category Feature Description References
New feature SQL plan management The SQL plan management feature is supported. This feature stores an execution plan for a set of SQL statements that share the same pattern and can run the execution plan with appropriate adjustments if an executed SQL statement shares the same pattern as the preceding statements. In addition, stored execution plans can be queried and revised manually. SQL plan management
Optimized feature Hint The hint feature is enabled by default. Use the hint feature
Shared input scan The shared input scan operator is optimized to improve the performance of highly concurrent complex common table expression (CTE) queries. None

February 21, 2022 (V6.3.8.0)

Category Feature Description References
New feature Query Profiling Statistics The query profiling statistics feature is supported to automatically collect and record statistics of AnalyticDB for PostgreSQL during query execution. Query statistics can be displayed in system views. Query profiling statistics
Dynamic Join Filter The dynamic join filter feature is supported. This feature can be used before a hash join to filter out the data that does not match the hash join from the left table by using the dynamically collected join key of the right table. This way, the disk reading, networking, and CPU computing overheads are reduced to improve the hash join performance. Dynamic join filter
Data synchronization to data warehouses The data synchronization feature is supported to synchronize data from ApsaraDB RDS for PostgreSQL to AnalyticDB for PostgreSQL. None
Support for Realtime Compute for Apache Flink Realtime Compute for Apache Flink can be used to write data to or read data from AnalyticDB for PostgreSQL based on the latest version of Flink AnalyticDB for PostgreSQL connector.
Optimized feature Error message for overlong fields An error message can be returned if the length of a written field exceeds the specified value when data is updated or inserted. None
Bitmap scan The index scan tuple statistics feature is supported for bitmap scans. The number of times for which indexes are used and the number of rows scanned can be returned if the execution plan of an SQL statement contains bitmap indexes. Manage indexes
Partitioned table The JIT cache feature is supported for partitioned tables to prevent repeated compilation. If more than 100 partitioned tables are involved, a query can be implemented within seconds. Define table partitioning
Laser computing engine The MergeAppend operator is supported for the Laser computing engine. If all subqueries are ordered and the ultimate results need to be sorted based on the same sort key, the MergeAppend operator can return a combination of ordered subquery results by using the heapsort algorithm. Use the Laser computing engine
Object Storage Service (OSS) foreign table Data of OSS foreign tables can be read and written in the Avro format. Use OSS foreign tables to access OSS data
Backup and restoration The parallel archiving feature is supported for backup and restoration to speed up incremental backup. Overview
Fixed issue The issue that may cause redistribution in a table that has a distribution key of the VARCHAR or TEXT type is fixed. Redistribution may be caused after operations such as UNION ALL are performed because the distribution key is missing.
The tuplestore backward scanning feature is supported to prevent Backward scanning of tuplestores are not supported at this time errors.
The issue that may cause DDL operation failures on secondary coordinator nodes during a switchover between the primary and standby coordinator nodes is fixed.

January 7, 2022 (V6.3.7.0)

Category Feature Description References
New feature pg_hint_plan extension The pg_hint_plan extension is supported. This extension uses hints to intervene in and optimize execution plans, and registers the optimized SQL patterns and hints. Then, optimized execution plans are automatically generated for SQL statements that use the same SQL patterns to improve execution efficiency.
Query cache The query cache is supported to cache query results. This feature allows duplicate queries to be executed within milliseconds. Query cache
Optimized feature LLVM JIT The LLVM JIT feature provided by PostgreSQL 14 is supported. None
Backup and restoration The parallel archiving feature is supported to improve backup performance. Back up data
Fixed issue The issue that may cause ineffective ORDER BY clauses in subqueries when the primary query contains only a LIMIT clause is fixed.
The issue that may cause rebalance failures during data recovery across compute nodes due to a missing mirror host in the pg_hba.conf file of the primary compute node is fixed.
The issue that may cause failure to create partitioned tables on the primary coordinator node in a multi-coordinator architecture is fixed.