You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by GitBox <gi...@apache.org> on 2020/05/31 16:30:46 UTC
[GitHub] [spark] wangyum commented on a change in pull request #28575: [SPARK-31705][SQL] Push predicate through join by rewriting join condition to conjunctive normal form
wangyum commented on a change in pull request #28575:
URL: https://github.com/apache/spark/pull/28575#discussion_r432962886
##########
File path: sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/FilterPushdownSuite.scala
##########
@@ -1230,4 +1233,154 @@ class FilterPushdownSuite extends PlanTest {
comparePlans(Optimize.execute(query.analyze), expected)
}
+
+ test("inner join: rewrite filter predicates to conjunctive normal form") {
+ val x = testRelation.subquery('x)
+ val y = testRelation.subquery('y)
+
+ val originalQuery = {
+ x.join(y)
+ .where(("x.b".attr === "y.b".attr)
+ && (("x.a".attr > 3) && ("y.a".attr > 13) || ("x.a".attr > 1) && ("y.a".attr > 11)))
+ }
+
+ val optimized = Optimize.execute(originalQuery.analyze)
+ val left = testRelation.where(('a > 3 || 'a > 1)).subquery('x)
+ val right = testRelation.where('a > 13 || 'a > 11).subquery('y)
+ val correctAnswer =
+ left.join(right, condition = Some("x.b".attr === "y.b".attr
+ && (("x.a".attr > 3) && ("y.a".attr > 13) || ("x.a".attr > 1) && ("y.a".attr > 11))))
+ .analyze
+
+ comparePlans(optimized, correctAnswer)
+ }
+
+ test("inner join: rewrite join predicates to conjunctive normal form") {
+ val x = testRelation.subquery('x)
+ val y = testRelation.subquery('y)
+
+ val originalQuery = {
+ x.join(y, condition = Some(("x.b".attr === "y.b".attr)
+ && (("x.a".attr > 3) && ("y.a".attr > 13) || ("x.a".attr > 1) && ("y.a".attr > 11))))
+ }
+
+ val optimized = Optimize.execute(originalQuery.analyze)
+ val left = testRelation.where('a > 3 || 'a > 1).subquery('x)
+ val right = testRelation.where('a > 13 || 'a > 11).subquery('y)
+ val correctAnswer =
+ left.join(right, condition = Some("x.b".attr === "y.b".attr
+ && (("x.a".attr > 3) && ("y.a".attr > 13) || ("x.a".attr > 1) && ("y.a".attr > 11))))
+ .analyze
+
+ comparePlans(optimized, correctAnswer)
+ }
+
+ test("inner join: rewrite join predicates(with NOT predicate) to conjunctive normal form") {
+ val x = testRelation.subquery('x)
+ val y = testRelation.subquery('y)
+
+ val originalQuery = {
+ x.join(y, condition = Some(("x.b".attr === "y.b".attr)
+ && Not(("x.a".attr > 3)
+ && ("x.a".attr < 2 || ("y.a".attr > 13)) || ("x.a".attr > 1) && ("y.a".attr > 11))))
+ }
+
+ val optimized = Optimize.execute(originalQuery.analyze)
+ val left = testRelation.where('a <= 3 || 'a >= 2).subquery('x)
+ val right = testRelation.subquery('y)
+ val correctAnswer =
+ left.join(right, condition = Some("x.b".attr === "y.b".attr
+ && (("x.a".attr <= 3) || (("x.a".attr >= 2) && ("y.a".attr <= 13)))
+ && (("x.a".attr <= 1) || ("y.a".attr <= 11))))
+ .analyze
+ comparePlans(optimized, correctAnswer)
+ }
+
+ test("left join: rewrite join predicates to conjunctive normal form") {
+ val x = testRelation.subquery('x)
+ val y = testRelation.subquery('y)
+
+ val originalQuery = {
+ x.join(y, joinType = LeftOuter, condition = Some(("x.b".attr === "y.b".attr)
+ && (("x.a".attr > 3) && ("y.a".attr > 13) || ("x.a".attr > 1) && ("y.a".attr > 11))))
+ }
+
+ val optimized = Optimize.execute(originalQuery.analyze)
+ val left = testRelation.subquery('x)
+ val right = testRelation.where('a > 13 || 'a > 11).subquery('y)
+ val correctAnswer =
+ left.join(right, joinType = LeftOuter, condition = Some("x.b".attr === "y.b".attr
+ && (("x.a".attr > 3) && ("y.a".attr > 13) || ("x.a".attr > 1) && ("y.a".attr > 11))))
+ .analyze
+
+ comparePlans(optimized, correctAnswer)
+ }
+
+ test("right join: rewrite join predicates to conjunctive normal form") {
+ val x = testRelation.subquery('x)
+ val y = testRelation.subquery('y)
+
+ val originalQuery = {
+ x.join(y, joinType = RightOuter, condition = Some(("x.b".attr === "y.b".attr)
+ && (("x.a".attr > 3) && ("y.a".attr > 13) || ("x.a".attr > 1) && ("y.a".attr > 11))))
+ }
+
+ val optimized = Optimize.execute(originalQuery.analyze)
+ val left = testRelation.where('a > 3 || 'a > 1).subquery('x)
+ val right = testRelation.subquery('y)
+ val correctAnswer =
+ left.join(right, joinType = RightOuter, condition = Some("x.b".attr === "y.b".attr
+ && (("x.a".attr > 3) && ("y.a".attr > 13) || ("x.a".attr > 1) && ("y.a".attr > 11))))
+ .analyze
+
+ comparePlans(optimized, correctAnswer)
+ }
+
+ test("inner join: rewrite to conjunctive normal form avoid genereting too many predicates") {
+ val x = testRelation.subquery('x)
+ val y = testRelation.subquery('y)
+
+ val originalQuery = {
+ x.join(y, condition = Some(("x.b".attr === "y.b".attr)
+ && ((("x.a".attr > 3) && ("x.a".attr < 13) && ("y.c".attr <= 5))
+ || (("y.a".attr > 2) && ("y.c".attr < 1)))))
+ }
+
+ val optimized = Optimize.execute(originalQuery.analyze)
+ val left = testRelation.subquery('x)
+ val right = testRelation.where('c <= 5 || ('a > 2 && 'c < 1)).subquery('y)
+ val correctAnswer = left.join(right, condition = Some("x.b".attr === "y.b".attr
+ && ((("x.a".attr > 3) && ("x.a".attr < 13) && ("y.c".attr <= 5))
+ || (("y.a".attr > 2) && ("y.c".attr < 1))))).analyze
+
+ comparePlans(optimized, correctAnswer)
Review comment:
Match the `PostgreSQL`'s plan:
```
postgres=# explain select x.* from x join y on (x.b = y.b and ( (x.a > 3 and x.a < 13 and y.c <= 5) or (y.a > 2 and y.c < 1) ));
QUERY PLAN
-----------------------------------------------------------------------------------------
Merge Join (cost=207.30..402.86 rows=1927 width=16)
Merge Cond: (y.b = x.b)
Join Filter: (((x.a > 3) AND (x.a < 13) AND (y.c <= 5)) OR ((y.a > 2) AND (y.c < 1)))
-> Sort (cost=78.41..80.30 rows=754 width=12)
Sort Key: y.b
-> Seq Scan on y (cost=0.00..42.38 rows=754 width=12)
Filter: ((c <= 5) OR ((a > 2) AND (c < 1)))
-> Sort (cost=128.89..133.52 rows=1850 width=16)
Sort Key: x.b
-> Seq Scan on x (cost=0.00..28.50 rows=1850 width=16)
(10 rows)
```
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org