AjaxXu commented on a change in pull request #9027:
[FLINK-13161][table-blink-runtime]numBuckets calculate wrong in
BinaryHashBucketArea
URL: https://github.com/apache/flink/pull/9027#discussion_r301869889
##########
File path:
flink-table/flink-table-runtime-blink/src/main/java/org/apache/flink/table/runtime/hashtable/BinaryHashBucketArea.java
##########
@@ -203,7 +203,7 @@ private void resize(boolean spillingAllowed) throws
IOException {
int oldNumBuckets = numBuckets;
MemorySegment[] oldOverflowSegments = overflowSegments;
int newNumSegs = oldBuckets.length * 2;
- int newNumBuckets = MathUtils.roundDownToPowerOf2(newNumSegs <<
table.bucketsPerSegmentBits);
Review comment:
I had changed the check condition. Is that OK? But in my option, setting
numBuckets to power of 2 is for fast matching. If this strategy is not change,
the buckets.length and numBuckets are always power of 2. Otherwise, it will
result in wasting segment.
----------------------------------------------------------------
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]
With regards,
Apache Git Services
