loserwang1024 commented on code in PR #4387: URL: https://github.com/apache/flink-cdc/pull/4387#discussion_r3303585024
########## flink-cdc-connect/flink-cdc-source-connectors/flink-connector-postgres-cdc/src/main/java/org/apache/flink/cdc/connectors/postgres/source/reader/PostgresSourceStreamFetcher.java: ########## @@ -0,0 +1,58 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.flink.cdc.connectors.postgres.source.reader; + +import org.apache.flink.cdc.connectors.base.source.reader.external.FetchTask; +import org.apache.flink.cdc.connectors.base.source.reader.external.IncrementalSourceStreamFetcher; + +import io.debezium.data.Envelope; +import org.apache.kafka.connect.data.Struct; +import org.apache.kafka.connect.source.SourceRecord; + +/** + * Stream fetcher for Postgres. When {@code includeLogicalMessages} is enabled, lets {@code + * pg_logical_emit_message} records (op="m") bypass table-based watermark filtering, since logical + * messages are not bound to a table. + */ +public class PostgresSourceStreamFetcher extends IncrementalSourceStreamFetcher { Review Comment: I don't think this is a good choice to a new subclass PostgresSourceStreamFetcher here. Maybe we can move shouldEmit to FetchTask.Context(such as isDataChangeRecord). When we design context and dialect control the different behaviors for different databases. The IncrementalSourceStreamFetcher shares same logical. ########## flink-cdc-connect/flink-cdc-source-connectors/flink-connector-postgres-cdc/src/main/java/org/apache/flink/cdc/connectors/postgres/source/reader/PostgresSourceRecordEmitter.java: ########## @@ -43,6 +45,18 @@ public PostgresSourceRecordEmitter( offsetFactory); } + @Override + protected void processElement( + SourceRecord element, SourceOutput<T> output, SourceSplitState splitState) + throws Exception { + if (PostgresSourceStreamFetcher.isLogicalMessage(element)) { Review Comment: PostgresSourceStreamFetcher is not a util class. Please move to `PostgresTypeUtils` or `PostgresQueryUtils` or a new `PostgresSourceRecordUtils`. ########## flink-cdc-connect/flink-cdc-source-connectors/flink-connector-postgres-cdc/src/main/java/org/apache/flink/cdc/connectors/postgres/source/reader/PostgresSourceStreamFetcher.java: ########## @@ -0,0 +1,58 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.flink.cdc.connectors.postgres.source.reader; + +import org.apache.flink.cdc.connectors.base.source.reader.external.FetchTask; +import org.apache.flink.cdc.connectors.base.source.reader.external.IncrementalSourceStreamFetcher; + +import io.debezium.data.Envelope; +import org.apache.kafka.connect.data.Struct; +import org.apache.kafka.connect.source.SourceRecord; + +/** + * Stream fetcher for Postgres. When {@code includeLogicalMessages} is enabled, lets {@code + * pg_logical_emit_message} records (op="m") bypass table-based watermark filtering, since logical + * messages are not bound to a table. + */ +public class PostgresSourceStreamFetcher extends IncrementalSourceStreamFetcher { + + private final boolean includeLogicalMessages; + + public PostgresSourceStreamFetcher( + FetchTask.Context taskContext, int subtaskId, boolean includeLogicalMessages) { + super(taskContext, subtaskId); + this.includeLogicalMessages = includeLogicalMessages; + } + + @Override + protected boolean shouldEmit(SourceRecord sourceRecord) { + if (includeLogicalMessages && isLogicalMessage(sourceRecord)) { + return true; + } + return super.shouldEmit(sourceRecord); + } + + static boolean isLogicalMessage(SourceRecord record) { + if (record.value() instanceof Struct) { + Struct struct = (Struct) record.value(); Review Comment: A minor confusion regarding PostgreSQL: The WAL logs in PostgreSQL are substituted based on table names. However, the logical messages being read now seem to lack identity identifiers, making it impossible to distinguish between subscriptions (i.e., different jobs would all read the same data). I have searched how to use logical message, it seems that prefix is used for it. ```sql pg_logical_emit_message ( transactional boolean , prefix text , content text [, flush boolean DEFAULTfalse ] ) ``` Therefore, I do think we need to add a param to control which prefix is needed to filter the target logical message. It's the key of this issue, @leonardBang , WDYT? ########## flink-cdc-connect/flink-cdc-source-connectors/flink-connector-postgres-cdc/src/main/java/org/apache/flink/cdc/connectors/postgres/source/config/PostgresSourceOptions.java: ########## @@ -98,6 +98,16 @@ public class PostgresSourceOptions extends JdbcSourceOptions { + "(1) PUBLICATION must be created beforehand with parameter publish_via_partition_root=true\n" + "(2) Table list (regex or predefined list) should only match the parent table name, if table list matches both parent and child tables, snapshot data will be read twice."); + public static final ConfigOption<Boolean> SCAN_LOGICAL_MESSAGE_ENABLED = + ConfigOptions.key("scan.logical-message.enabled") Review Comment: we need also to tell the use in this config description and document: 1. which version of postgres supports messages in pgoutput. 2. how to enable it. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
