ilooner commented on a change in pull request #1606: DRILL-6845: Semi-Hash-Join
to skip incoming build duplicates, automatically stop skipping if too few
URL: https://github.com/apache/drill/pull/1606#discussion_r248149275
##########
File path:
exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/join/HashJoinBatch.java
##########
@@ -973,9 +973,32 @@ public IterOutcome executeBuildPhase() throws
SchemaChangeException {
// Do auto tuning
buildCalc = partitionNumTuning(maxBatchRowCount, buildCalc);
}
- // to be used in case of a Semi Join skippinging duplicates
- spillControlCalc = new HashJoinSpillControlImpl(allocator,
RECORDS_PER_BATCH,
- (int)
context.getOptions().getOption(ExecConstants.HASHJOIN_MIN_BATCHES_IN_AVAILABLE_MEMORY_VALIDATOR),
batchMemoryManager);
+ if ( semiSkipDuplicates ) {
+ // in case of a Semi Join skippinging duplicates, use a "spill
control" calc
+ // (may revert back to the buildCalc if the code decides to stop
skipping)
+ currentCalc = new HashJoinSpillControlImpl(allocator,
RECORDS_PER_BATCH,
+ (int)
context.getOptions().getOption(ExecConstants.HASHJOIN_MIN_BATCHES_IN_AVAILABLE_MEMORY_VALIDATOR),
Review comment:
I understand we want to have a knob to allow for some slack. But there is
already the safety factor parameter that does this. Why give the user yet
another tuning parameter when one already exists? It makes things more
difficult for the user.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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
