Gengliang Wang created SPARK-57025:
--------------------------------------
Summary: SortMergeJoinExec: extract JoinHelper.resetMatched for
full-outer BitSet bookkeeping
Key: SPARK-57025
URL: https://issues.apache.org/jira/browse/SPARK-57025
Project: Spark
Issue Type: Sub-task
Components: SQL
Affects Versions: 5.0.0
Reporter: Gengliang Wang
{{SortMergeJoinExec.codegenFullOuter}} and its interpreted counterpart
{{SortMergeFullOuterJoinScanner.findMatchingRows}} both duplicate the "reuse
the BitSet if its capacity is enough, otherwise allocate a new one" idiom for
tracking matched left/right rows.
Extract it into a static helper
{code:java}
JoinHelper.resetMatched(BitSet matched, int bufferSize): BitSet
{code}
and call it from the four sites (left + right in each method). Replaces ~16
inline lines with 4 helper calls and keeps the codegen and interpreted paths in
lockstep.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
