Re: Why RexUtil.toCnf() has version that takes a limit but RexUtil.toDnf() doesn't ?

Thanks! I'll create a Jira for this and PR accordingly. Yingyu Wang On Fri, Aug 12, 2022 at 7:00 PM Julian Hyde wrote: > Git blame reveals that that line of code was added when fixing > https://issues.apache.org/jira/browse/CALCITE-1290 < > https://issues.apache.org/jira/browse/CALCITE-1290>.

Re: Why RexUtil.toCnf() has version that takes a limit but RexUtil.toDnf() doesn't ?

Git blame reveals that that line of code was added when fixing https://issues.apache.org/jira/browse/CALCITE-1290 . https://github.com/apache/calcite/commit/77315c06 . I see no reason why

Why RexUtil.toCnf() has version that takes a limit but RexUtil.toDnf() doesn't ?

Does anyone know why RexUtil.toCnf() has a version of the method that takes a limit arg: toCnf(RexBuilder rexBuilder, int maxCnfNodeCount, RexNode rex), but there is no such version for RexUtil.toDnf()? Is there any special reason for this? Can we add a similar version for RexUtil.toDnf() then?