zhangbutao commented on PR #34775: URL: https://github.com/apache/doris/pull/34775#issuecomment-2112527866
> If you want to do this optimization, why not just create a dummy IcebergSplit? So we even don't need to call `TableScanUtil.planTasks` @morningman Thanks for your suggestion! You are right, create a dummy IcebergSplit is a better approach than this PR. But i found that BE need a real iceberg spilt to do some code logic, and we need do some odd check in BE side to let the BE accept the dummy IcebergSplit. Acutually, i think the `count `pushdown is a metadata operation which can be done in fe side, and so we can skip to initialize the split and no need the operation in BE side. This can be improve increase efficiency of `count `operation. I just submitted a new PR https://github.com/apache/doris/pull/34928, Please take a look if you have time. Thanks. -- 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: commits-unsubscr...@doris.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org For additional commands, e-mail: commits-h...@doris.apache.org
