b08cf6e822
### What changes were proposed in this pull request? This PR improves `Repartition` and `RepartitionByExpr` statistics estimation using child statistics. ### Why are the changes needed? The current implementation will missing column stat. For example: ```sql CREATE TABLE t1 USING parquet AS SELECT id % 10 AS key FROM range(100); ANALYZE TABLE t1 COMPUTE STATISTICS FOR ALL COLUMNS; set spark.sql.cbo.enabled=true; EXPLAIN COST SELECT key FROM (SELECT key FROM t1 DISTRIBUTE BY key) t GROUP BY key; ``` Before this PR: ``` == Optimized Logical Plan == Aggregate [key#2950L], [key#2950L], Statistics(sizeInBytes=1600.0 B) +- RepartitionByExpression [key#2950L], Statistics(sizeInBytes=1600.0 B, rowCount=100) +- Relation default.t1[key#2950L] parquet, Statistics(sizeInBytes=1600.0 B, rowCount=100) ``` After this PR: ``` == Optimized Logical Plan == Aggregate [key#2950L], [key#2950L], Statistics(sizeInBytes=160.0 B, rowCount=10) +- RepartitionByExpression [key#2950L], Statistics(sizeInBytes=1600.0 B, rowCount=100) +- Relation default.t1[key#2950L] parquet, Statistics(sizeInBytes=1600.0 B, rowCount=100) ``` ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? Unit test. Closes #32309 from wangyum/SPARK-35203. Authored-by: Yuming Wang <yumwang@ebay.com> Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org> |
||
---|---|---|
.. | ||
benchmarks | ||
src | ||
pom.xml |