Yunhong Zheng created FLINK-30971: ------------------------------------- Summary: Modify the default value of parameter 'table.exec.local-hash-agg.adaptive.sampling-threshold' Key: FLINK-30971 URL: https://issues.apache.org/jira/browse/FLINK-30971 Project: Flink Issue Type: Bug Components: Table SQL / Runtime Affects Versions: 1.17.0 Reporter: Yunhong Zheng Fix For: 1.17.0
In our test environment, we set the default parallelism to 1 and got the most appropriate default value of parameter 'table.exec.local-hash-agg.adaptive.sampling-threshold' is 5000000. However, for these batch jobs with high parallelism in produce environment, the amount of data in single parallelism is almost less than 5000000. Therefore, after testing, we found that set to 500000 can get better results. -- This message was sent by Atlassian Jira (v8.20.10#820010)