PostgreSQL agent changelog

Fixed an issue that could have caused duplicated PKs to be displayed in the UI if a column was both a primary key and a unique key;

Fixed join conditions that can cause duplicated rows error;

Improved performance of the agent during bulk loads;

Updated vendor libraries.

Fixed column names escaping for Bulk Operations;

Fixed primary keys columns management in staging table creation during bulk operations;

Added support for Bulk inserts, both while performing a Snapshot task and while performing a CDC task (For snapshot tasks, it is enabled by default, for CDC tasks it is disabled by default);

GSSD-374: Fixed deserialization of M messages incoming from PostgreSQL WAL stream version 2, with null xid;

GSSD-374: Fixed deserialization of M messages incoming from PostgreSQL WAL stream version 2;

GSSD-374: Added support for custom enum data types;

Replaced INSERT …​. ON CONFLICT statement with INSERT + UPDATE;

Added support for unique key;

Fixed a bug on date types parsing from CDC stream;

Fixed an issue with datetime parsing;

GSSD-322: Fixed a regression that could have caused the agent to fail storing data in the target database in the case of CDC and logical keys enabled;

Minor bugfixes and improvements.

Forced commit after truncate operation ahead of snapshot task.

Improved error handling;

Added support for XML and JSON data types;

Minor bugfixes and improvements.

Support for quoted identifiers;

Fixed an issue that could have caused the agent to ingest single-row transactions with a reduced rate;

Added support for additional data types:

Fixed an issue that was preventing the TRUNCATE operation to be committed;

Fixed an issue when reading timestamp(n) data types from PostgreSQL replication slots;

Fixed an issue when reading timestamp(n) data types from PostgreSQL while a snapshot is in progress;