b-slim commented on a change in pull request #1386:
URL: https://github.com/apache/samza/pull/1386#discussion_r443023108
##########
File path:
samza-sql/src/main/java/org/apache/samza/sql/planner/QueryPlanner.java
##########
@@ -141,20 +142,28 @@ public RelRoot plan(String query) {
.sqlToRelConverterConfig(SqlToRelConverter.Config.DEFAULT)
.traitDefs(traitDefs)
.context(Contexts.EMPTY_CONTEXT)
- .costFactory(null)
+ .programs(
+ Programs.hep(ImmutableList.of(FilterJoinRule.FILTER_ON_JOIN),
true, DefaultRelMetadataProvider.INSTANCE))
Review comment:
@atoomula this rule is not really the main part of this work and I don't
think we need to fix/choose now which rule should be used to handle filters or
should or should not be pushed to table scan this is beyond the scope of this
PR.
The main goal is to explore how we can handle the **operator stack between
the join and remote table scan**.
I added the rule to ensure that if a filter is there things works as
expected and it seems it is working. But I would love to know what are the
**blind spot(s)** I am missing about what work and does not work well with the
remote table scan if you think this will help drive this issue. Thanks
----------------------------------------------------------------
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:
[email protected]
