[ https://issues.apache.org/jira/browse/FLINK-23531?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
tartarus updated FLINK-23531: ----------------------------- Description: current we keep all change log for row-time deduplicate mini-batch, we need buffer all record on heap. In some cases, users do not need all change logs; In this way, the memory pressure and the calculation pressure of downstream operators can be reduced was: current we keep everytime change log for row-time deduplicate mini-batch, we need buffer all record on heap. In some cases, users do not need all change logs; In this way, the memory pressure and the calculation pressure of downstream operators can be reduced > Allow skip all change log for row-time deduplicate mini-batch > ------------------------------------------------------------- > > Key: FLINK-23531 > URL: https://issues.apache.org/jira/browse/FLINK-23531 > Project: Flink > Issue Type: Improvement > Components: Table SQL / Runtime > Reporter: tartarus > Priority: Major > > current we keep all change log for row-time deduplicate mini-batch, > we need buffer all record on heap. > In some cases, users do not need all change logs; > In this way, the memory pressure and the calculation pressure of downstream > operators can be reduced -- This message was sent by Atlassian Jira (v8.3.4#803005)