[BlockJoin] Add ParentsChildrenBlockJoinQuery to support parent and c… #14728
Conversation
github-project-automation
bot
moved this to Open
in OpenSearch Lucene & Core Performance Tracking
|
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one. If the PR doesn't need a changelog entry, then add the skip-changelog-check label to it and you will stop receiving this reminder on future updates to the PR. |
github-project-automation
bot
moved this from Open
to Closed
in OpenSearch Lucene & Core Performance Tracking
github-project-automation
bot
moved this from Closed
to Open
in OpenSearch Lucene & Core Performance Tracking
|
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one. If the PR doesn't need a changelog entry, then add the skip-changelog-check label to it and you will stop receiving this reminder on future updates to the PR. |
lucene/join/src/test/org/apache/lucene/search/join/TestParentsChildrenBlockJoinQuery.java
Outdated
Show resolved
Hide resolved
lucene/join/src/test/org/apache/lucene/search/join/TestParentsChildrenBlockJoinQuery.java
Outdated
Show resolved
Hide resolved
lucene/join/src/test/org/apache/lucene/search/join/TestParentsChildrenBlockJoinQuery.java
Outdated
Show resolved
Hide resolved
lucene/join/src/java/org/apache/lucene/search/join/ParentsChildrenBlockJoinQuery.java
Outdated
Show resolved
Hide resolved
Jinny-Wang
force-pushed
the
jinnyw_block_join_new_query
branch
from
03117a9 to
edbd182
Compare
|
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one. If the PR doesn't need a changelog entry, then add the skip-changelog-check label to it and you will stop receiving this reminder on future updates to the PR. |
…hild docs filtering in one pass
Jinny-Wang
force-pushed
the
jinnyw_block_join_new_query
branch
from
edbd182 to
5132163
Compare
|
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one. If the PR doesn't need a changelog entry, then add the skip-changelog label to it and you will stop receiving this reminder on future updates to the PR. |
lucene/join/src/java/org/apache/lucene/search/join/ParentsChildrenBlockJoinQuery.java
Show resolved
Hide resolved
|
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one. If the PR doesn't need a changelog entry, then add the skip-changelog label to it and you will stop receiving this reminder on future updates to the PR. |
|
@mkhludnev -- you had some comments on the linked issue. Does this approach seem reasonable to you? If so, I'll go ahead and merge it. Also, we can safely backport this to 10.x. @Jinny-Wang, could you please move the changelog entry to 10.3.0? Thanks! |
|
@msfroh I still think extending the existing ToChildBlockJoinQuery with this functionality is preferable, at least from a code maintenance standpoint. However, I can't contribute much to this work now, or anytime soon. |
github-project-automation
bot
moved this from Open
to Merged
in OpenSearch Lucene & Core Performance Tracking
#14565
Problem
The current block join query family doesn't support filtering on both parent docs and child docs at the same time while applying a childLimitPerParent.
To do that we either need to use a ToChildBlockJoinQuery combined with another childQuery.
Or use ToParentBlockJoinQuery combined with ParentChildrenBlockJoinQuery (which only takes 1 parent doc id).
More detailed examples and description in #14565
Solution
Introduce a new
ParentsChildrenBlockJoinQueryso that it supports one pass scanning on both parent and child docs , applying filter to both parent and child at the same time while supporting a child limit per parent.