I see two failures due to checkerframework. Those are real. Let us know when you've fixed them.
The failure in Druid is probably due to a Druid bug that existed in November. You can ignore it. Should go away after rebase. Please keep nagging on this thread... this is a good change, and we should get it into 1.37. On Mon, Jan 8, 2024 at 11:34 AM Mihai Budiu <mbu...@gmail.com> wrote: > > CI failed, so perhaps you can look into that. > > Mihai > ________________________________ > From: Paul Jackson <pauljackson...@verizon.net.INVALID> > Sent: Monday, January 8, 2024 7:06 AM > To: dev@calcite.apache.org <dev@calcite.apache.org> > Subject: Re: PR 3495 (Make single-row rels unique) Review Request > > Is there anything else to do for this PR? > > * https://github.com/apache/calcite/pull/3495 > * https://issues.apache.org/jira/browse/CALCITE-6044 > > Thanks, > -Paul > > > > On Nov 16, 2023, at 7:51 AM, Paul Jackson > > <pauljackson...@verizon.net.INVALID> wrote: > > > > This PR updates org.apache.calcite.rel.metadata areColumnsUnique, > > getUniqueKeys, and getPredicates methods to account for relations that > > produce at most a single row: Sort (due to limit), Aggregate (due to no > > group by), Values (due to having a single row), and Join (when joining with > > one of the other nodes). > > https://github.com/apache/calcite/pull/3495https://issues.apache.org/jira/projects/CALCITE/issues/CALCITE-6044 > > Thanks,-Paul