Michael Smith has posted comments on this change. ( http://gerrit.cloudera.org:8080/21729 )
Change subject: IMPALA-13185: Include runtime filter source in key ...................................................................... Patch Set 6: (10 comments) http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/main/java/org/apache/impala/planner/PlanNode.java File fe/src/main/java/org/apache/impala/planner/PlanNode.java: http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/main/java/org/apache/impala/planner/PlanNode.java@544 PS5, Line 544: for (Expr e: conjuncts_) { > I played with that and had some problems. I'll revisit it. Done http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/main/java/org/apache/impala/planner/PlanNode.java@1362 PS5, Line 1362: "Build-side contains current node, > Here are some ways that I test by hand (with Impala running with tuple cach Added a test covering this. http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/main/java/org/apache/impala/planner/PlanNode.java@1364 PS5, Line 1364: if (build instanceof ExchangeNode && build.getChildCount() == > TFileSplitGeneratorSpec has enough information, but that information is enc Done http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/main/java/org/apache/impala/planner/RuntimeFilterGenerator.java File fe/src/main/java/org/apache/impala/planner/RuntimeFilterGenerator.java: http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/main/java/org/apache/impala/planner/RuntimeFilterGenerator.java@395 PS5, Line 395: } else { > Nit: We might want to skip this for tuple caching. Done http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/main/java/org/apache/impala/planner/RuntimeFilterGenerator.java@403 PS5, Line 403: } : } > I would appreciate a few words saying that the list of targets includes oth Done http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/test/java/org/apache/impala/planner/TupleCacheTest.java File fe/src/test/java/org/apache/impala/planner/TupleCacheTest.java: http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/test/java/org/apache/impala/planner/TupleCacheTest.java@69 PS5, Line 69: public void testRuntimeFilterCacheKeys() { > Something that can help with tests that use joins is to use straight_join t That seems like I should add straight_join to every example? http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/test/java/org/apache/impala/planner/TupleCacheTest.java@81 PS5, Line 81: verifyDifferentCacheKeys(basicJoin, "select a.id from functional.alltypes a, " + : "functional.alltypes b where a.id = b.id"); > In addition to this test case, let's add one where the build side has an ad Done, I think. http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/test/java/org/apache/impala/planner/TupleCacheTest.java@111 PS5, Line 111: */ > From an eligibility point of view, we can add some tests where a Kudu table Done http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/test/java/org/apache/impala/planner/TupleCacheTest.java@139 PS5, Line 139: /** > There's a way we can add a test for this. Basically, that query that has a I don't think that works because the probe side is ineligible for caching because the build is too complicated. I had a couple verifyOverlappingCacheKeys cases looking at generality in testRuntimeFilterCacheKeys; should I move them here? http://gerrit.cloudera.org:8080/#/c/21729/5/fe/src/test/java/org/apache/impala/planner/TupleCacheTest.java@371 PS5, Line 371: "default", System.getProp > One quirk of this test so far is that it uses num_nodes=1, so there are no I think I've added that in test_tuple_cache.py. -- To view, visit http://gerrit.cloudera.org:8080/21729 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I0077964be5acdb588d76251a6a39e57a0f42bb5a Gerrit-Change-Number: 21729 Gerrit-PatchSet: 6 Gerrit-Owner: Michael Smith <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Joe McDonnell <[email protected]> Gerrit-Reviewer: Kurt Deschler <[email protected]> Gerrit-Reviewer: Michael Smith <[email protected]> Gerrit-Reviewer: Steve Carlin <[email protected]> Gerrit-Reviewer: Yida Wu <[email protected]> Gerrit-Comment-Date: Wed, 25 Sep 2024 23:21:35 +0000 Gerrit-HasComments: Yes
