[FLINK-38840][postgres] Postgres YAML connector supports emitting complete Table ID

[Mrart]

Currently, Postgres YAML connector maps (schema, table) as Pipeline event Table ID, and the database part is missing.

We may add a new option to generate complete Table ID (database, schema, table). This could be useful when records are written into Kafka as JSON messages.

[Mrart]

@yuxiqian PTAL

[loserwang1024]

@Mrart
It seems that even we don't have table-id.include-database option but directly emitting complete Table ID. The compability of currently pipeline won't be broke. And kafka sink can add a new option to determine whether to print database. @yuxiqian , WDYT?

[yuxiqian]

@Mrart It seems that even we don't have table-id.include-database option but directly emitting complete Table ID. The compability of currently pipeline won't be broke. And kafka sink can add a new option to determine whether to print database. @yuxiqian , WDYT?

I'm a little concerned about changing default Table ID format. It may affect transform & route table matching rules and schema registry states where Table IDs are used as key.

[Mrart]

@Mrart It seems that even we don't have table-id.include-database option but directly emitting complete Table ID. The compability of currently pipeline won't be broke. And kafka sink can add a new option to determine whether to print database. @yuxiqian , WDYT?

We still need to control the output of the full tableID through configuration to ensure that it is compatible with production

[yuxiqian]

Changes LGTM. Would @loserwang1024 like to take another look?

[Mrart]

@loserwang1024 PTLK

[loserwang1024]

LGTM