kennknowles opened a new issue, #19562: URL: https://github.com/apache/beam/issues/19562
When Calcite is planning the query it may ask the table provider for the same table multiple times and in different alternative query plans there might be different instances of the same table. Since the row count estimation is stored in the table instances, each time that the row count estimation is called for a new instance, the table creates a new estimate. The estimation may take some time; therefore, this can potentially downgrade the performance (of planning) and increase the planning time. There are two potential ways to solve this problem: 1- Make sure that the table providers do not create multiple instances for the same table. or 2- Keep the row count estimations in a common data structure or a static Map and reuse it in multiple instances when needed. Imported from Jira [BEAM-7655](https://issues.apache.org/jira/browse/BEAM-7655). Original Jira may contain additional context. Reported by: riazela. -- 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]
