2021

Category

Feature

Description

References

New feature

sql mapping

SQL statement mapping is supported for PolarDB for PostgreSQL (Compatible with Oracle). Incorrect SQL statements can be mapped to correct SQL statements. This way, issues caused by incorrect SQL statements can be prevented.

polar_sql_mapping

Support for parallel update in cross-node parallel execution

Parallel update is supported in cross-node parallel execution, including data update that contains subqueries.

You can set the polar_px_update_dop_num parameter to control the concurrent number of parallel update. Default value: 6. Valid values: 1 to 128.

Cross-node parallel execution

Support for parallel delete in cross-node parallel execution

Parallel delete is supported in cross-node parallel execution, including parallel delete that contains subqueries and parallel delete that does not contain subqueries.

You can set the polar_px_delete_dop_num parameter to control the concurrent number of parallel delete. Default value: 6. Valid values: 1 to 128.

Support for partition-wise join in cross-node parallel execution

The advanced feature partition-wise join is supported in cross-node parallel execution.

Query data from partitioned tables

Compatibility with Oracle views

PolarDB for PostgreSQL (Compatible with Oracle) is compatible with the all_tab_columns view, the all_tab_comments view, and some columns of the all_objects view.

ALL_TAB_COLUMNS and ALL_OBJECTS

Optimized feature

Support of DBMS_SQL.TO_CURSOR_NUMBER and DBMS_SQL.TO_REFCURSOR

The DBMS_SQL.TO_CURSOR stored procedure is supported.

   the conversion from local dynamic SQL to DBMS_SQL.
  curid := dbms_sql.to_cursor_number(cursrc);

None

Category

Feature

Description

References

New feature

ST_PixelAsPolygon function

The ST_PixelAsPolygon function is supported. This function is used to convert pixels to polygon objects.

ST_PixelAsPolygon

ST_PixelAsPolygons function

The ST_PixelAsPolygons function is supported. This function is used to convert pixels to collections of polygon objects.

ST_PixelAsPolygons

ST_PixelAsPoint function

The ST_PixelAsPoint function is supported. This function is used to convert pixels to geometric objects expressed in points.

ST_PixelAsPoint

ST_PixelAsCentroid function

The ST_PixelAsCentroid function is supported. This function is used to convert pixels to geometric objects expressed around a center point.

ST_PixelAsCentroid

ST_PixelAsPoints function

The ST_PixelAsCentroid is supported. This function is used to convert pixels to collections of point objects.

ST_PixelAsPoints

ST_PixelAsCentroids function

The ST_PixelAsCentroids function is supported. This function is used to convert pixels to collections of center point objects.

ST_PixelAsCentroids

Optimized feature

Geometry model

Geometry is compatible with the PostGIS V3.1.4.

None

ST_AsDatasetFile function

The ST_AsDatasetFile function is supported. This function is used to return data in the GRIB2 format.

None

Fixed issue

The following issue is fixed: The obtained special reference is incorrect when NetCDF data is imported to a raster object.

The following issue is fixed: An error occurs during the update of multiple bounding boxes due to the incorrect result of the equals method.

Category

Feature

Description

References

New feature

Partition pruning supported on partitioned tables at runtime for DML operations

You can execute the EXPLAIN statement when you update or delete a partitioned table to prune partitions at runtime.

Partition pruning

ORDER BY clause supported in aggregate functions

When you execute an aggregate function, you can use ORDER BY for simple queries.

Aggregate functions

Support of schemas specified for triggers

You can specify a schema that precedes the name of a trigger for the trigger.

Trigger

Support for two-dimensional associative arrays

Two-dimensional associative arrays are supported.

Associative arrays

Support for multiple read and multiple write to accelerate queries with INSERT INTO SELECT in cross-node parallel execution

In addition to multiple read and single write implemented by executing INSERT INTO SELECT, cross-node parallel execution also supports parallel write and parallel read to further improve the performance of INSERT INTO SELECT.

Cross-node parallel execution

Support for the acceleration of tableless queries that use INSERT INTO SELECT in cross-node parallel execution

Cross-node parallel execution supports the acceleration of tableless queries that use INSERT INTO SELECT, which further improves the performance of INSERT INTO SELECT.

Support for forced order in INSERT INTO SELECT for order-sensitive SQL queries in cross-node parallel execution

Cross-node parallel execution allows you to set polar_px_enable_insert_order_sensitive to ensure that parallel write operations are performed in order. Default value: true.

• If polar_px_enable_insert_order_sensitive is set to true and an SQL statement contains ORDER BY or the tableless query feature is used, parallel write operations are forced to perform in order.

• If polar_px_enable_insert_order_sensitive is set to false, parallel write operations are not necessarily performed in order.

Support for ROWNUM in cross-node parallel execution

You can set the ROWNUM expression that is compatible with Oracle for cross-node parallel execution.

Interruption support for automatic interruption of cross-node parallel queries after a timeout for the continuity of conflicted DDL operations

To prevent DDL operations from undergoing sustained suspension due to lock conflicts with cross-node parallel queries, you can set the duration for which the operations have to wait. If the lockout duration ends, cross-node parallel queries are forced to interrupt and the operations continue.

Support for LEADING hints

You can configure more than one table for LEADING hints to specify the join order of multiple tables.

LEADING hints

DELETE FROM subquery

Support of DELETE FROM for subqueries

None

Support for SELECT UNIQUE

The SELECT UNIQUE syntax is supported. It is equivalent to SELECT DISTINCT.

None

UPDATING IN SPL TRIGGER

In DML triggers, the UPDATING ('colname') conditional predicate is supported.

None

Support of WHERE for UPDATE and DELETE in the MERGE statement

In the MERGE statement, you can use WHERE with UPDATE or DELETE.

None

Support of aliases for INSERT INTO SELECT

You can set alias for INSERT INTO SELECT.

None

Support of delimiters between characters for COPY

You can use delimiters to separate multiple characters for COPY.

None

Optimized feature

Support for polar_row_id in cross-node parallel execution

You can set the polar_row_id parameter for cross-node parallel execution.

None

Parallel playback for read-only nodes

Playback is accelerated for read-only nodes by using multithreading, which reduces the latency of synchronization between the primary node and read-only nodes.

None

Category

Feature

Description

References

Optimized feature

ST_ImportFrom function

The ST_ImportFrom function is enhanced to allow Object Storage Service (OSS) objects to be mapped to files in memory and accelerate data import.

ST_ImportFrom and ganos.raster.memory_oss_file_max_size

GiST and TrajGisT indexing

The trajgist_ops_multi operator family is supported for Generalized Search Tree (GiST) and TrajGisT indexing. You can create a GiST or TrajGisT index on multiple bounding boxes.

GiST indexing, TrajGiST indexing, and ganos.trajectory.index_split_config

Fixed issue

Vector pyramids are optimized to support queries of metadata.

The issue that causes memory leaks when Ganos is used to evaluate whether two specified trajectories intersect is fixed.

Category

Feature

Description

References

New feature

Support for using statement outlines to rewrite queries

Statement outlines can be used to rewrite SQL queries when the query rewrite feature of PolarDB cannot meet your business requirements.

Adjust execution plans

Support for Assert operators in cross-node parallel execution

The Assert operator is supported by the cross-node parallel execution feature to check runtime constraints on the executor.

Cross-node parallel execution

Support for Oracle DECODE expressions in cross-node parallel execution

Oracle DECODE expressions are supported by the cross-node parallel execution feature.

Dynamic partition pruning in cross-node parallel execution

Dynamic partition pruning is supported by the cross-node parallel execution feature to improve the performance of hash joins for partitioned tables.

Support for holdable portals in Bind messages

SQL statements can be accessed across transactions. To enable this feature, add polar_holdable_hint hints to SQL statements. The following SQL statement is an example: SELECT /*+ polar_holdable_hint */;.

None

SELECT FOR UPDATE statement compatible with Oracle syntax

The SELECT FOR UPDATE OF statement is fully compatible with Oracle syntax.

None

Optimized feature

Support for constant expressions in cross-node parallel execution

Constant expressions are supported by the cross-node parallel execution feature.

None

Operator-level memory limits

Operator-level memory limits can be used to reduce out-of-memory (OOM) risks and improve PolarDB stability.

None

Category

Feature

Description

References

Optimized feature

Functions that are used to compute raster surfaces

The functions that are used to compute raster surfaces are optimized. The optimized functions return raster objects instead of pixel matrices.

ST_Aspect, ST_Slope, and ST_Hillshade

ST_Split function

The ST_Split function is optimized. The optimized function splits a trajectory object based on a specified rule and returns an array of sub-trajectories.

ST_Split

Fixed issue

The following issue is fixed: The ST_SrEqual function fails to determine whether two spatial reference systems are the same in specific scenarios.

Category

Feature

Description

References

New feature

Support for using the Index Advisor to log execution costs before and after indexing

Two columns are added to the index_advisor_log table to indicate the execution costs that are incurred before and after each index is created. This helps upper-layer applications such as Database Autonomy Service (DAS) evaluate the benefits of indexing.

None

Accelerated creation of global indexes by cross-node parallel execution

Cross-node parallel execution can be used to accelerate the process of creating global B-tree indexes.

Cross-node parallel execution, Accelerate index creation, and Create a global index

Category

Feature

Description

References

New feature

Compatibility with GNU Compiler Collection (GCC) 9

PolarDB for PostgreSQL (Compatible with Oracle) is compatible with GCC 9.2.1. This can optimize compilation and improve performance.

None

Hints in statement outlines

Hints can be specified in statement outlines to control how the optimizer generates execution plans based on the statement outlines.

Adjust execution plans

Cross-node parallel execution on hash partitions

Parallel execution across hash partitions is supported. This helps balance I/O loads and improve system performance.

Query data from partitioned tables

DDL operations on partitioned tables that have global indexes defined

Global indexes on a partitioned table can be managed when you perform DDL operations such as DROP, SPLIT, and MERGE on the partitioned table.

Supported DDL operations on partitioned tables

Optimized feature

Compatibility with the ORA_HASH function

The ORA_HASH function is replaced with the DBMS_UTILITY.GET_HASH_VALUE function. The DBMS_UTILITY.GET_HASH_VALUE function is used to calculate the hash value of a specified string.

None

Category

Feature

Description

References

New feature

ST_RemoveDriftPoints function

The ST_RemoveDriftPoints function is supported. This function is used to delete drift points from a trajectory based on specified rules.

ST_removeDriftPoints

ST_Split function

The ST_Split function is supported. This function is used to split a trajectory based on a specified geometry object.

ST_Split

ST_ExportTo function

The ST_ExportTo function is supported. This function is used to export a trajectory from a database to an external file for storage. Only metadata is retained in the database.

ST_ExportTo

ST_IsExternal function

The ST_IsExternal function is supported. This function is used to check whether a trajectory is stored in an external file.

ST_IsExternal

ST_ImportFrom function

The ST_ImportFrom function is supported. This function is used to import a trajectory from an external file back to a database.

ST_importFrom

ST_StorageLocation function

The ST_StorageLocation function is supported. This function is used to return the external location of a trajectory.

ST_StorageLocation

ST_AKID function

The ST_AKID function is supported. This function is used to return an AccessKey ID that is used to read a trajectory from OSS.

ST_AKID

ST_SetAccessKey function

The ST_SetAccessKey function is supported. This function is used to configure an AccessKey pair that is used to read a trajectory from OSS. An AccessKey pair consists of an AccessKey ID and an AccessKey secret. An AccessKey ID is similar to a logon username, and an AccessKey secret is similar to a password.

ST_SetAccessKey

ST_SetAKID function

The ST_SetAKID function is supported. This function is used to configure an AccessKey ID that is used to read a trajectory from OSS.

ST_SetAkId

ST_SetAKSecret function

The ST_SetAKSecret function is supported. This function is used to configure an AccessKey secret that is used to read a trajectory from OSS.

ST_SetAkSecret

ST_SetStorageLocation function

The ST_SetStorageLocation function is supported. This function is used to configure the external location from which a trajectory can be read.

ST_SetStorageLocation

ST_DeleteGtf function

The ST_DeleteGtf function is supported. This function is used to delete all exported trajectory files in a specified folder.

ST_DeleteGTF

Fixed issue

The following issue is fixed: Data cannot be filtered by strings when you create a vector pyramid.

The following issue is fixed: The layer list cannot be displayed when QGIS is connected.

Category

Feature

Description

References

New feature

Display of I/O statistics in the polar_stat_sql and polar_stat_plan views

• The relevant resource information can be obtained by querying different views. This can help identify performance issues.

• Performance diagnostics is supported by PolarDB for PostgreSQL (Compatible with Oracle). This allows you to analyze performance issues related to instances and SQL queries.

• Relevant parameters are supported to display I/O statistics in the polar_stat_sql and polar_stat_plan views.

Performance views and Performance diagnostics

Cross-node parallel execution

Cross-node parallel execution can be used to scan Block Range Indexes (BRIN indexes). This further improves query performance.

Overview of cross-node parallel execution

The polar_bt_write_page_buffer_size parameter is supported. This parameter specifies the write I/O policy that is used while an index is being created.

Accelerate index creation

The following three granularities can be used to specify a query range when you use cross-node parallel execution:

• System: specifies whether cross-node parallel execution is applied to all the queries that run in all sessions.

• Session: specifies whether cross-node parallel execution is applied to the current session.

• Query: specifies whether cross-node parallel execution is applied to specific queries.

Run analytical queries

Cross-node parallel execution can be used for partitioned tables to improve database performance.

Query data from partitioned tables

Optimized feature

polar_crash_recovery_rto parameter

The polar_crash_recovery_rto parameter is supported. This parameter specifies an expected recovery time objective (RTO) for an instance. This ensures that databases on an instance can restart for recovery and become available within the expected RTO if the instance fails.

None

Category

Feature

Description

References

New feature

ST_AsDatasetFile function

The ST_AsDatasetFile function is supported. This function is used to return a specified range of raster objects as a binary stream in the specified file format.

ST_AsDatasetFile

ST_RasterDrivers function

The ST_RasterDrivers function is supported. This function is used to query the status of all raster data drivers.

ST_RasterDrivers

Fixed issue

The following issue is fixed: A reprojection fails when you use the ST_Clip function to specify a spatial reference system for the reprojection.

Category

Feature

Description

References

New feature

Cross-node parallel execution

Cross-node parallel execution is supported.

• SQL queries can be run across multiple read-only nodes in a distributed manner. This improves the query performance of PolarDB for PostgreSQL (Compatible with Oracle).

• Cross-node parallel execution can be used to accelerate the process of creating indexes in scenarios where large tables are used.

Cross-node parallel execution

Support of partitioned tables for global indexes

Global indexes can be created on partitioned tables. Indexes can be created in parallel. This means that DML operations on a partitioned table are not blocked while an index is being created on the table.

Create a global index

Optimized feature

Performance monitoring

The max_slot_wal_keep_size parameter is supported. This parameter is used to limit the size of files that use replication slots. This can prevent Write-Ahead Logging (WAL) files from accumulating.

None

DBLink

By default, functions that do not meet collation requirements and ANY() and ALL() expressions can be pushed down. This improves query performance. You can use the polar_enable_pushable_unsafe_collate_remote and polar_enable_pushable_all_any_remote parameters to disable these settings.

Use a database link to query data across databases

Rownum

ROWNUM is optimized. As a result, ROWNUM values can be compared with the values of the VARCHAR2 or TEXT type.

None

polar_resource_group plug-in

Multiple parameters including total_mem_limit_remain_size, idle_mem_limit_rate, enable_terminate_active, policy_mem_release, and enable_log can be used to configure the polar_resource_group plug-in. This avoids excessive usage of memory and ensures the normal running of database processes.

None

Improved startup speed of read-only nodes

The startup of read-only nodes is sped up. This improves performance.

None

Failover

The parallel playback feature can be used to accelerate the recovery of database services after a failover. In most cases, the database services can be recovered within 15 seconds after a primary database fails.

None

Category

Feature

Description

References

New feature

Spatio-temporal parallel query

Spatial indexes can be queried in parallel. This improves query efficiency.

Enable the spatio-temporal parallel query feature

Optimized feature

Raster image processing

The ST_SubRaster function can be used to convert the pixel types of raster objects and stretch the pixel values. Raster image processing can be used in AI recognition scenarios in which multi-band remote sensing images are converted to three-band images.

ST_SubRaster

Fixed issue

The following issue is fixed: A database may unexpectedly exit when you perform a raster data analysis on invalid histogram information.

The following issue is fixed: After a specified raster dataset is transformed into another projection, a failure may occur when you disconnect from your database.

The following issue is fixed: A database may unexpectedly exit when you import raster data to the database.

Category

Feature

Description

References

New feature

idle_session_timeout parameter

The idle_session_timeout parameter is supported. This parameter specifies the period during which sessions can remain idle before the sessions time out. Sessions are terminated after they time out.

• Valid values: 0 to 2147483647

  The default value of 0 specifies that the timeout feature is disabled for idle sessions.

• Unit

  • 'ms': millisecond. This is the default unit.

  • 's': second.

  • 'min': minute.

  • 'h': hour.

  • 'd': day.

Configure parameters

DDL operations related to constraints

The ALTER TABLE ... ENABLE/DISABLE CONSTRAINT SQL statement is supported. This statement can be used to enable or disable constraints.

ALTER TABLE ... ENABLE/DISABLE CONSTRAINT

DDL operations related to tablespaces

DDL operations related to tablespaces are supported. For example, CREATE TABLESPACE can be executed to create tablespaces.

None

Numeric data types

This minor version is compatible with the numeric data types of Oracle. You can enable this feature by specifying the polar_enable_comp_numeric_format parameter. By default, this feature is disabled. To enable this feature, execute the following statement:

SET polar_enable_comp_numeric_format=on;

None

Fixed issue

The following issue is fixed: The built-in DBMS_JOB package cannot be used across databases.

The following issue is fixed: The CONNECT BY clause and the ROWNUM function cannot be used together.

The following issue is fixed: The CONNECT BY clause cannot find parent objects.

Category

Feature

Description

References

New feature

Raster data type

The ST_RPCGeoreference function is supported. This function is used to retrieve the information about rational polynomial coefficients (RPCs) of a raster object.

ST_RPCGeoreference

The ST_SetRPCGeoreference function is supported. This function is used to specify RPC parameters for a raster object.

ST_SetRPCGeoreference

The ST_RPCRectify function is supported. This function is used to rectify a raster object based on the RPC parameters and return the rectified raster object.

ST_RPCRectify

GiST indexing

GiST indexes can be created in parallel.

Create spatial indexes in parallel

Fixed issue

The following issue is fixed: When you use the pg_dump utility, custom spatial references cannot be exported.

Category

Feature

Description

References

New feature

Partitioned table

BEFORE row-level triggers can be used in partitioned tables.

BEFORE row-level trigger

polar_outline plug-in

The polar_outline plug-in can be used to adjust execution plans. This optimizes SQL statements.

Adjust execution plans

Optimized feature

Partitioned table

Subpartition templates can be used for interval range partitioning.

Subpartition templates

Performance improvements:

The I/O performance of reading data files can be improved by caching information about table sizes into shared memory.

None

Compatibility

The following operators are supported by INTERVAL and FLOAT8 data types:

• +

• -

• *

• /

None

Compatibility

The OPT_PARAM hint is added. You can use this hint to specify the values of specific parameters in a single SQL statement.

OPT_PARAM hints

Category

Feature

Description

References

New feature

Storage optimization for large spatio-temporal objects

The Simple storage policy is supported. This policy optimizes storage for large spatio-temporal objects.

Use the Simple storage policy

Optimized feature

Raster object

The chunked data in a raster object can be automatically deleted when the raster object is updated by using a new raster object.

None

Fixed issue

The following issue is fixed: The Ganos Trajectory extension fails to be upgraded.

The following issue is fixed: In some cases, an error occurs when a raster object is resampled by using the average resampling method.

The following issue is fixed: If the trajectory points for multiple timestamps are the same in a trajectory object, query results may be invalid.

The following issue is fixed: The conversion process is terminated after a GEOS object fails to be converted. After this issue is fixed, the system also prompts the causes of the conversion failure.

Category

Feature

Description

References

New feature

Partitioned table

Interval range partitioning is supported. Interval range partitioning is an extension to range partitioning. Interval range partitioning allows a database to automatically create a partition if data to be inserted exceeds the range of an existing partition.

Interval range partitioning

Range-partitioned tables can be converted into interval-range partitioned tables. Interval-range partitioned tables can be converted into range-partitioned tables.

Switch between range partitioning and interval range partitioning

Subpartition templates are supported. You can use subpartition templates to define subpartitions and simplify the specifications of subpartitions. You need to define subpartition descriptors in a subpartition template only once and then apply the subpartition template to each partition in a table. This way, you can define the subpartition descriptors for multiple partitions in the table at a time.

Subpartition templates

Automatic Workload Repository (AWR) for performance monitoring

AWR is supported. AWR collects and analyzes performance statistics. You can use AWR to generate an analytical report for a specified period based on statistics such as dynamic views of a database. This allows you to analyze the database performance within the specified period.

None

Optimized feature

Partitioned table

Partition pruning is enabled in the execution plan for a partitioned table. This improves query performance.

None

Data type

PolarDB for PostgreSQL (Compatible with Oracle) uses TIMESTAMP as the DATE data type. A TIMESTAMP value is accurate to microseconds. In Oracle, a DATE value is accurate to seconds. To ensure compatibility between PolarDB for PostgreSQL (Compatible with Oracle) and Oracle, the system automatically converts a DATE value into a TIMESTAMP(0) value.

None

WAL

The pipeline technology is supported. This technology is used to reduce the proportion of WAL overhead. This can improve overall system performance.

None

TO_CHAR

TO_CHAR(text1, text2) is supported. This function automatically identifies the data type of text1 and text2 on the server side. If text1 and text2 are valid DATE values, the function automatically converts text1 into a TIMESTAMP value.

None

Fixed issue

Invalid SIGHUP signals are deleted.

The following issue is fixed: The system stops responding when you execute the MERGE INTO statement for a partitioned table.

The following issue is fixed: An "illegal memory access" error occurs when you use pg_icu_collate_names.

The following issue is fixed: An "illegal memory access" error occurs when you execute the ALTER QUEUE ADD CALL statement.

The following issue is fixed: An "illegal memory access" error occurs when you enter a string of a variable length in the rdname field of a POLICY system table. This error is returned because the data type of the rdname field is set to a fixed-length data type.

The following issue is fixed: The return values of functions are truncated because the functions are not declared in the Transparent Data Encryption (TDE) feature.

The following issue is fixed: The roaringbitmap plug-in encounters an "illegal memory access" error when invalid parameters are specified.

The following issue is fixed: If the synchronization latency is high when logs are synchronized from a primary node to read-only nodes, the read-only nodes may not immediately provide feedback on streaming replication to the primary node. As a result, the streaming replication is interrupted.

Category

Feature

Description

References

New feature

Storage optimization for large spatio-temporal objects

The feature signatures of large spatio-temporal objects are used. This accelerates the process of creating indexes on spatio-temporal data, makes spatial queries more efficient, and optimizes the storage modes of spatio-temporal objects. The polar_enable_storage_partial parameter is also supported. You can specify this parameter to enable the metadata of spatio-temporal objects to be stored in-row and the entity data of spatio-temporal objects to be stored off-row.

Feature signature-based storage optimization for large spatio-temporal objects

Parallel execution of functions supported by the raster module

The ST_ImportFrom function is supported. This function is used to import raster data in parallel.

ST_ImportFrom

The ST_BuildPyramid function is supported. This function is used to create pyramids in parallel based on raster data.

ST_BuildPyramid

Trajectory Empty

The Trajectory Empty object is supported. This allows null objects to be returned if requirements are not met in some scenarios.

None

Fixed issue

The following issue is fixed: When the input parameter is a 16-bit signed integer (16BSI), the ST_AddZ function may return an invalid result.

The following issue is fixed: In some cases, the trajectory module fails to be upgraded.

The following issue is fixed: Some functions cannot use indexes in the trajectory module.

The following issue is fixed: An error instead of the value false is returned by some nd functions. This issue may occur if these nd functions fail to resolve the problem that two time periods do not intersect.

Category

Feature

Description

References

New feature

DBMS_STATS plug-in

The DBMS_STATS plug-in is supported. You can use this plug-in to collect relevant statistics while the database is running, back up statistics, and restore statistics from backups.

DBMS_STATS

Transaction splitting after a write request

Transaction splitting is supported for transactions that contain statements for read requests that are sent after a write request at the READ COMMITTED isolation level. All read requests that are generated from the start time to the end time of a transaction can be sent to read-only nodes. This ensures read/write consistency in the transaction.

Transaction splitting

Optimized feature

Support for expressions that contain the OR operator

In earlier versions, an expression that contains the OR operator cannot be converted into an expression that contains the UNION ALL operator. This makes queries inefficient because full table scans are performed. In the latest version, an expression that contains the OR operator is automatically converted into an expression that contains the UNION ALL operator. This optimizes the logic and improves query performance.

None

Compatibility with ECPG

• ECPGPlus is compatible with the partial syntax of Pro*C.

• ECPGPlus is compatible with the connection methods that are supported by native Oracle.

None

Compatibility with Oracle

The GROUP BY CONST/EMPTY clause is compatible with Oracle syntax.

None

PBP

By default, the persistent buffer pool for a primary node is enabled.

None

Synchronization of partitioned tables to Kafka

The process of synchronizing the configurations of a partitioned table from PolarDB for PostgreSQL (Compatible with Oracle) to Kafka is optimized. Data is synchronized to Kafka based on parent tables. After data is synchronized, parent tables that use the same names as the source tables are created on Kafka.

None

Users

The following syntax is adjusted: CREATE USER ... IDENTIFIED BY .... This syntax is used to create only standard users.

Create a user

Fixed issue

The following issue is fixed: When logical replication is used to synchronize DBMS_JOB tables that do not have primary keys, tasks fail to be deleted.

The following issue is fixed: A memory leak occurs in the processes of resource groups if the memory usage exceeds the upper limit.

The following issue is fixed: Standby Promote requires a long period of time.

The following issue is fixed: The temporary table parameter in the temporary field of the ***_all_tables view is abnormally displayed.

The following issue is fixed: A "memory access out of bounds" error is recorded in audit logs.

The following issue is fixed: When you use a database link that connects a PolarDB for PostgreSQL (Compatible with Oracle) database to a PostgreSQL database to query data, the query result is truncated.

The issues that occur during subpartition pruning based on multi-partition key columns for partitioned tables are fixed.

Category

Feature

Description

References

New feature

ST_JaccardSimilarity function

The ST_JaccardSimilarity function is supported. This function is used to calculate the similarity between trajectory objects.

ST_JaccardSimilarity

ST_Transform function

The ST_Transform function is supported. This function is used to transform the coordinate of a trajectory between spatial reference systems.

ST_Transform

Optimized feature

Vector pyramid

The spatial extent can be specified when you create a vector pyramid. This fixes the issue that the spatial extent of a vector pyramid cannot be automatically calculated in some scenarios.

None

Fixed issue

The following issue is fixed: Ganos data catalogs are misconfigured in some environments.