impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tim Armstrong (Code Review)" <ger...@cloudera.org>
Subject [Impala-ASF-CR] IMPALA-4866: Hash join node does not apply limits correctly
Date Tue, 23 May 2017 21:13:26 GMT
Tim Armstrong has posted comments on this change.

Change subject: IMPALA-4866: Hash join node does not apply limits correctly
......................................................................


Patch Set 2:

(6 comments)

Looking pretty good I think.

http://gerrit.cloudera.org:8080/#/c/6778/2/be/src/exec/partitioned-hash-join-node.cc
File be/src/exec/partitioned-hash-join-node.cc:

Line 582:       COUNTER_SET(rows_returned_counter_, num_rows_returned_);
Maybe we should set the counter at the bottom of GetNext(), instead of in the multiple places
that it is currently set (this relates to my comment about decrementing 'num_rows_returned_'
when we truncate the batch).


Line 642:     out_batch->set_num_rows(out_batch->num_rows() - (num_rows_returned_ -
limit_));
There's a bug if the hash join node is in a subplan - SubplanNode may call into HashJoinNode
with a partially-filled batch. In that case the number of rows returned is already accounted
in 'num_rows_returned_'.


Line 643:     *eos = true;
Shouldn't we decrement 'num_rows_returned_' if we truncated the batch? Otherwise it won't
be accurate in the profile.


Line 806:     num_rows_returned_ += 1;
Nit: we usually write this as:

  ++num_rows_returned_;


Line 944:       num_rows_returned_ += 1;
Nit: we usually write this as:

  ++num_rows_returned_;


http://gerrit.cloudera.org:8080/#/c/6778/2/testdata/workloads/functional-query/queries/QueryTest/single-node-joins-with-limits.test
File testdata/workloads/functional-query/queries/QueryTest/single-node-joins-with-limits.test:

How long do these take to run? Should they be under exhaustive?


-- 
To view, visit http://gerrit.cloudera.org:8080/6778
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I414124f8bb6f8b2af2df468e1c23418d05a0e29f
Gerrit-PatchSet: 2
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: anujphadke <aphadke@cloudera.com>
Gerrit-Reviewer: Dan Hecht <dhecht@cloudera.com>
Gerrit-Reviewer: Matthew Jacobs <mj@cloudera.com>
Gerrit-Reviewer: Tim Armstrong <tarmstrong@cloudera.com>
Gerrit-Reviewer: anujphadke <aphadke@cloudera.com>
Gerrit-HasComments: Yes

Mime
View raw message