I don’t know. It’s possible that with the current join-ordering algorithm the query has a much better plan. Can you try, and post the current plan to the JIRA case.
If it’s still an issue, fixing it is probably not that easy. The problem might be in the statistics, or in the join-ordering algorithm, or a few other places. It’s probably not a bug where you can easily point to the line of code that is wrong. Julian > On May 19, 2018, at 9:57 AM, Atri Sharma <[email protected]> wrote: > > Hi All, > > Is CALCITE - 99 still applicable? Is it worth a shot? > > https://issues.apache.org/jira/browse/CALCITE-99 > > Regards, > > Atri
