Michael Ho has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11535 )

Change subject: IMPALA-6661 Make NaN values equal for grouping purposes.
......................................................................


Patch Set 16:

(3 comments)

Hey sorry for not getting back earlier. Bogged down again with some users' 
issues which I have to help with.

http://gerrit.cloudera.org:8080/#/c/11535/16/testdata/workloads/functional-query/queries/QueryTest/aggregation.test
File testdata/workloads/functional-query/queries/QueryTest/aggregation.test:

http://gerrit.cloudera.org:8080/#/c/11535/16/testdata/workloads/functional-query/queries/QueryTest/aggregation.test@1384
PS16, Line 1384: (VALUES((1.6 x, 2 y), (3.2, 4), (5.4,6))) T
Please see comments in joins.test


http://gerrit.cloudera.org:8080/#/c/11535/16/testdata/workloads/functional-query/queries/QueryTest/joins.test
File testdata/workloads/functional-query/queries/QueryTest/joins.test:

http://gerrit.cloudera.org:8080/#/c/11535/16/testdata/workloads/functional-query/queries/QueryTest/joins.test@801
PS16, Line 801: (VALUES((1.6 x, 0 y), (3.2, 1), (5.4,2), (0.5, 3), (0.5, 4), 
(-0.5, 5))) XX),
The problem with this kind of query with few number VALUES() is that codegen 
will be disabled as the planner knows the number of rows will be small. I think 
it may make sense to have another test cases to scan some sizable table. This 
is also a good test case to keep as this exercises the interpretation path.

Of course, one can also set the query option DISABLE_CODEGEN_ROWS_THRESHOLD to 
a small value but it seems better to have a more realistic test query with 
scans in there instead of joining two union nodes of constants.

You can check the query profile to see codegen is enabled in the HASH JOIN node.


http://gerrit.cloudera.org:8080/#/c/11535/16/testdata/workloads/functional-query/queries/QueryTest/joins.test@818
PS16, Line 818: with q as (VALUES((cast(1.0 as FLOAT) x), (2.0))),
              :      r as (select t1.x from q t1, q t2 where sqrt(1.0-t1.x) <=> 
sqrt(1.0-t2.x))
              :      select * from r
If you plan to keep this test case, this can be simplified as:
  with q as (VALUES((cast(1.0 as FLOAT) x), (2.0)))
      select t1.x from q t1, q t2 where sqrt(1.0-t1.x) <=> sqrt(1.0-t2.x)



--
To view, visit http://gerrit.cloudera.org:8080/11535
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I996c4a2e1934fd887046ed0c55457b7285375086
Gerrit-Change-Number: 11535
Gerrit-PatchSet: 16
Gerrit-Owner: Michal Ostrowski <mostr...@cloudera.com>
Gerrit-Reviewer: Bikramjeet Vig <bikramjeet....@cloudera.com>
Gerrit-Reviewer: Impala Public Jenkins <impala-public-jenk...@cloudera.com>
Gerrit-Reviewer: Michael Ho <k...@cloudera.com>
Gerrit-Reviewer: Michal Ostrowski <mostr...@cloudera.com>
Gerrit-Reviewer: Paul Rogers <par0...@yahoo.com>
Gerrit-Reviewer: Thomas Marshall <thomasmarsh...@cmu.edu>
Gerrit-Reviewer: Tim Armstrong <tarmstr...@cloudera.com>
Gerrit-Comment-Date: Fri, 19 Oct 2018 18:18:15 +0000
Gerrit-HasComments: Yes

Reply via email to