spark-instrumented-optimizer/sql/core
tanel.kiis@gmail.com bb6d6b5602 [SPARK-33964][SQL] Combine distinct unions in more cases
### What changes were proposed in this pull request?

Added the `RemoveNoopOperators` rule to optimization batch `Union`.  Also made sure that the `RemoveNoopOperators` would be idempotent.

### Why are the changes needed?

In several TPCDS queries the `CombineUnions` rule does not manage to combine unions, because they have noop `Project`s between them.
The `Project`s will be removed by `RemoveNoopOperators`, but by then `ReplaceDistinctWithAggregate` has been applied and there are aggregates between the unions. Adding a copy of `RemoveNoopOperators` earlier in the optimization chain allows `CombineUnions` to work on more queries.

### Does this PR introduce _any_ user-facing change?

No

### How was this patch tested?

New UTs and the output of `PlanStabilitySuite`

Closes #30996 from tanelk/SPARK-33964_combine_unions.

Authored-by: tanel.kiis@gmail.com <tanel.kiis@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
2021-01-05 11:01:31 +09:00
..
benchmarks [SPARK-33523][SQL][TEST][FOLLOWUP] Fix benchmark case name in SubExprEliminationBenchmark 2020-11-25 15:22:47 -08:00
src [SPARK-33964][SQL] Combine distinct unions in more cases 2021-01-05 11:01:31 +09:00
pom.xml [SPARK-33662][BUILD] Setting version to 3.2.0-SNAPSHOT 2020-12-04 14:10:42 -08:00