Uh oh!
There was an error while loading. Please reload this page.
[SPARK-28885][SQL][FOLLOW-UP] Re-enable the ported PgSQL regression tests of SQLQueryTestSuite - #26492
[SPARK-28885][SQL][FOLLOW-UP] Re-enable the ported PgSQL regression tests of SQLQueryTestSuite #26492maropu wants to merge 4 commits into
Conversation
SparkQA
commented
Nov 13, 2019
Test build #113669 has finished for PR 26492 at commit
|
maropu
commented
Nov 13, 2019
retest this please |
maropu
commented
Nov 13, 2019
Thanks for your quick check, @MaxGekk ! cc: @dongjoon-hyun@gengliangwang |
SparkQA
commented
Nov 13, 2019
Test build #113682 has finished for PR 26492 at commit
|
| INSERT INTO fract_only VALUES (5, '0.99994'); | ||
| -- INSERT INTO fract_only VALUES (3, '1.0' as decimal(4,4)); -- should fail | ||
| INSERT INTO fract_only VALUES (4, cast('-0.9999' as decimal(4,4))); | ||
| INSERT INTO fract_only VALUES (5, cast('0.99994' as decimal(4,4))); |
There was a problem hiding this comment.
So, this is converting CASTING DURING INSERTING into CASTING WITH CAST syntax.
There was a problem hiding this comment.
Ur... my fault.. I'll update later.
There was a problem hiding this comment.
Oh, it was just a comment to clarify. I think it's okay for now AS-IS since we don't support the original query in any way.
There was a problem hiding this comment.
But, the original queries throw unsupported exceptions now, so most of queries in this file become invalid in Spark. In the last commit, I changed these string literals into floating-point literals for depending on INSERT assignment casts. Which do you think is better, as-is or the last commit?
| INSERT INTO TIMESTAMP_TBL VALUES ('now'); | ||
| -- PostgreSQL implicitly casts string literals to data with timestamp types, but | ||
| -- Spark does not support that kind of implicit casts. | ||
| INSERT INTO TIMESTAMP_TBL VALUES timestamp(('now')); |
There was a problem hiding this comment.
It seems that you wanted the following.
- INSERT INTO TIMESTAMP_TBL VALUES timestamp(('now'));
+ INSERT INTO TIMESTAMP_TBL VALUES timestamp('now');
Or,
INSERT INTO TIMESTAMP_TBL VALUES (timestamp('now'));
SparkQA
commented
Nov 14, 2019
Test build #113751 has finished for PR 26492 at commit
|
gengliangwang
commented
Nov 14, 2019
@maropu Thanks for reenabling the test cases! |
dongjoon-hyun
commented
Nov 14, 2019
| -- PostgreSQL implicitly casts string literals to data with timestamp types, but | ||
| -- Spark does not support that kind of implicit casts. | ||
| INSERT INTO TIMESTAMP_TBL VALUES timestamp(('now')); | ||
| INSERT INTO TIMESTAMP_TBL VALUES (timestamp('now')); |
There was a problem hiding this comment.
Did we regenerate timestamp.sql.out?
| -- !query 1 | ||
| INSERT INTO TIMESTAMP_TBL VALUES ('now') | ||
| INSERT INTO TIMESTAMP_TBL VALUES timestamp(('now')) |
SparkQA
commented
Nov 14, 2019
Test build #113759 has finished for PR 26492 at commit
|
gengliangwang
commented
Nov 14, 2019
retest this please. |
SparkQA
commented
Nov 14, 2019
Test build #113766 has finished for PR 26492 at commit
|
maropu
commented
Nov 15, 2019
retest this please |
SparkQA
commented
Nov 15, 2019
Test build #113823 has finished for PR 26492 at commit
|
maropu
commented
Nov 20, 2019
retest this please |
SparkQA
commented
Nov 20, 2019
Test build #114124 has finished for PR 26492 at commit
|
maropu
commented
Nov 20, 2019
Ur, it seems I need to update the golden file again.. |
202988b to
7c06f0dCompareSparkQA
commented
Nov 20, 2019
Test build #114154 has finished for PR 26492 at commit
|
dongjoon-hyun
left a comment
There was a problem hiding this comment.
+1, LGTM. Merged to master.
maropu
commented
Nov 21, 2019
Thanks for taking care of it, guys! @dongjoon-hyun@gengliangwang@MaxGekk |
What changes were proposed in this pull request?
SPARK-28885(#26107) has supported the ANSI store assignment rules and stopped running some ported PgSQL regression tests that violate the rules. To re-activate these tests, this pr is to modify them for passing tests with the rules.
Why are the changes needed?
To make the test coverage better.
Does this PR introduce any user-facing change?
No.
How was this patch tested?
Existing tests.