You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@druid.apache.org by GitBox <gi...@apache.org> on 2020/04/25 08:18:54 UTC
[GitHub] [druid] maytasm commented on a change in pull request #9648: SQL: More straightforward handling of join planning.
maytasm commented on a change in pull request #9648:
URL: https://github.com/apache/druid/pull/9648#discussion_r415012376
##########
File path: sql/src/main/java/org/apache/druid/sql/calcite/rel/DruidJoinQueryRel.java
##########
@@ -395,4 +365,14 @@ private static boolean computeRightRequiresSubquery(final DruidRel<?> right)
return Pair.of(rightPrefix, signatureBuilder.build());
}
+
+ private static DruidRel<?> getSomeDruidChild(final RelNode child)
+ {
+ if (child instanceof DruidRel) {
+ return (DruidRel<?>) child;
+ } else {
+ final RelSubset subset = (RelSubset) child;
Review comment:
subset of equivalent rels -> Does this imply that the cost of each RelNode will be equal? I saw that computeLeftRequiresSubquery does returns different depending on which child we pick from the list of RelList. This then result in a very very different cost since a child that results in requiring subquery will have very high cost and a child that doesnt will have a much lower cost.
More specifically, I saw two RelNode in the list. One RelNode has a filter = null and the other has "filter":{"type":"selector","dimension":"v","value":"xa","extractionFn":null}
----------------------------------------------------------------
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: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org