[
https://issues.apache.org/jira/browse/IMPALA-5956?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16670606#comment-16670606
]
Tim Armstrong commented on IMPALA-5956:
---------------------------------------
Root cause of the q31 issue seems to be the same problem for a different query
as IMPALA-5950 - the query doesn't have the right substitution variables. The
order by doesn't include all the returned columns and ordering by the wrong
column turns out to not provide a total order on the rows. This is the diff to
switch to the standard qualification parameters.---
a/testdata/workloads/tpcds-unmodified/queries/tpcds-q31.test
+++ b/testdata/workloads/tpcds-unmodified/queries/tpcds-q31.test
@@ -29,30 +29,79 @@ with ss as
,ws ws3
where
ss1.d_qoy = 1
- and ss1.d_year = 2002
+ and ss1.d_year = 2000
and ss1.ca_county = ss2.ca_county
and ss2.d_qoy = 2
- and ss2.d_year = 2002
+ and ss2.d_year = 2000
and ss2.ca_county = ss3.ca_county
and ss3.d_qoy = 3
- and ss3.d_year = 2002
+ and ss3.d_year = 2000
and ss1.ca_county = ws1.ca_county
and ws1.d_qoy = 1
- and ws1.d_year = 2002
+ and ws1.d_year = 2000
and ws1.ca_county = ws2.ca_county
and ws2.d_qoy = 2
- and ws2.d_year = 2002
+ and ws2.d_year = 2000
and ws1.ca_county = ws3.ca_county
and ws3.d_qoy = 3
- and ws3.d_year =2002
+ and ws3.d_year =2000
and case when ws1.web_sales > 0 then ws2.web_sales/ws1.web_sales else null
end
> case when ss1.store_sales > 0 then ss2.store_sales/ss1.store_sales
else null end
and case when ws2.web_sales > 0 then ws3.web_sales/ws2.web_sales else null
end
> case when ss2.store_sales > 0 then ss3.store_sales/ss2.store_sales
else null end
- order by ss1.d_year;
-
-
+ order by ss1.ca_county;
{noformat}
--- a/testdata/workloads/tpcds-unmodified/queries/tpcds-q31.test
+++ b/testdata/workloads/tpcds-unmodified/queries/tpcds-q31.test
@@ -29,30 +29,79 @@ with ss as
,ws ws3
where
ss1.d_qoy = 1
- and ss1.d_year = 2002
+ and ss1.d_year = 2000
and ss1.ca_county = ss2.ca_county
and ss2.d_qoy = 2
- and ss2.d_year = 2002
+ and ss2.d_year = 2000
and ss2.ca_county = ss3.ca_county
and ss3.d_qoy = 3
- and ss3.d_year = 2002
+ and ss3.d_year = 2000
and ss1.ca_county = ws1.ca_county
and ws1.d_qoy = 1
- and ws1.d_year = 2002
+ and ws1.d_year = 2000
and ws1.ca_county = ws2.ca_county
and ws2.d_qoy = 2
- and ws2.d_year = 2002
+ and ws2.d_year = 2000
and ws1.ca_county = ws3.ca_county
and ws3.d_qoy = 3
- and ws3.d_year =2002
+ and ws3.d_year =2000
and case when ws1.web_sales > 0 then ws2.web_sales/ws1.web_sales else null
end
> case when ss1.store_sales > 0 then ss2.store_sales/ss1.store_sales
else null end
and case when ws2.web_sales > 0 then ws3.web_sales/ws2.web_sales else null
end
> case when ss2.store_sales > 0 then ss3.store_sales/ss2.store_sales
else null end
- order by ss1.d_year;
-
-
+ order by ss1.ca_county;
{noformat}
I.e. this isn't a product bug.
> Fix substitution variables and TPC-DS q31 to test suite
> -------------------------------------------------------
>
> Key: IMPALA-5956
> URL: https://issues.apache.org/jira/browse/IMPALA-5956
> Project: IMPALA
> Issue Type: Bug
> Components: Infrastructure
> Reporter: Tim Wood
> Assignee: Tim Armstrong
> Priority: Major
> Attachments: q59-flap.out, q89-flap1.out, q89-flap2.out, ttq-243.out,
> ttq-256.out
>
>
> When run esp. as part of the TPC-DS suite, query #89 returns varying results
> in the LSD of a calculation. Using the output of the previous run as the
> expected result for the next run fails. This SELECT item is a ROUND()
> expression, so it's not clear why the result is not deterministic.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
