[
https://issues.apache.org/jira/browse/FLINK-37479?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17946341#comment-17946341
]
Hongshun Wang edited comment on FLINK-37479 at 4/22/25 8:33 AM:
----------------------------------------------------------------
> I recall that if we use option publish_via_partition_root, it will return
> event as parent table and we shouldn't concern about the leaf table at all?
[~phamvinh1712] I am not sure about it. Someone used to tell me that when
publish_via_partition_root , the snapshot data can be read while stream split
is not read. However, I didn't test it by myself.
Maybe you can try it by set debezium.publication.autocreate.mode[1] as none,
and do publish with publish_via_partition_root by yourself to test? And then
use this slot in flink cdc to test latest mode? If can read latest data, it'll
be okay.
[1]
[{{publication.autocreate.mode}}|https://debezium.io/documentation/reference/3.1/connectors/postgresql.html#postgresql-publication-autocreate-mode]
was (Author: JIRAUSER298968):
> I recall that if we use option publish_via_partition_root, it will return
> event as parent table and we shouldn't concern about the leaf table at all?
[~phamvinh1712] I am not sure it. Maybe you can try it by set
debezium.publication.autocreate.mode[1] as none, and do publish with
publish_via_partition_root by yourself to test? And then use this slot in flink
cdc to test latest mode? If can read latest data, it'll be okay.
[1]
[{{publication.autocreate.mode}}|https://debezium.io/documentation/reference/3.1/connectors/postgresql.html#postgresql-publication-autocreate-mode]
> postgres cdc connector support discover PARTITIONED TABLE
> ---------------------------------------------------------
>
> Key: FLINK-37479
> URL: https://issues.apache.org/jira/browse/FLINK-37479
> Project: Flink
> Issue Type: Improvement
> Components: Flink CDC
> Reporter: Vinh Pham
> Priority: Minor
>
> At the moment, Postgresql connector doesn't support discover PARTITIONED
> TABLE, making the connector not possible to take snapshot on partitioned
> table.
> [https://github.com/apache/flink-cdc/blob/master/flink-cdc-connect/flink-cdc-source-connectors/flink-connector-postgres-cdc/src/main/java/org/apache/flink/cdc/connectors/postgres/source/utils/TableDiscoveryUtils.java#L41]
>
> This is inconsistent with the other implementation in PostgresConnection
> class
> [https://github.com/apache/flink-cdc/blob/master/flink-cdc-connect/flink-cdc-source-connectors/flink-connector-postgres-cdc/src/main/java/io/debezium/connector/postgresql/connection/PostgresConnection.java#L859]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
