Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
115304: opt: make floating-point stats calculations precise r=DrewKimball a=DrewKimball This patch makes two changes to the floating-point operators in the `adjustNullCountsForOuterJoins` statistics method: 1. The results of floating-point subtractions are adjusted to be at least `epsilon`; this avoids precision errors. 2. Calculation is avoided entirely when the result is known to be zero. This ensures an exact result in the common case when no adjustment is made to the null count estimate. This commit also reverts #115287, since the failing tests are now fixed. Fixes #115278 Release note: None Co-authored-by: Drew Kimball <drewk@cockroachlabs.com>
- Loading branch information