Github user eowhadi commented on a diff in the pull request:
https://github.com/apache/incubator-trafodion/pull/395#discussion_r56862361
--- Diff: core/sql/executor/ExHbaseSelect.cpp ---
@@ -444,6 +444,7 @@ ExWorkProcRetcode ExHbaseScanSQTaskTcb::work(short &rc)
Lng32 retcode = 0;
rc = 0;
Lng32 remainingInBatch = batchSize_;
+ NABoolean isFirstBatch = false;
--- End diff --
OK so here this boolean is not needed for the full length of the operator.
Only for the first 8K rows.
I can change the logic and maintain a row counter on the tcb. That will
have impact memory write for the full length of the operator. By doing my
initial trick, I was hopping that I could leverage the existing
remainingInBatch counter and not create another one. But I guess if I cannot
trust that the parent queue can fill up with less than an HBase_Buffer_size, I
cannot use the optimization.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
