Details
-
Type:
Bug
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: Qserv
-
Labels:
-
Story Points:8
-
Epic Link:
-
Sprint:DB_S15_07, DB_S15_08, DB_W16_09
-
Team:Data Access and Database
Description
IN is broken - it only uses the first element from the list. Here is the proof:
select COUNT(*) AS N FROM qservTest_case01_qserv.Source
|
WHERE objectId=386950783579546;
|
+------+
|
| N |
|
+------+
|
| 56 |
|
+------+
|
1 row in set (0.10 sec)
|
|
|
mysql> select count(*) AS N FROM qservTest_case01_qserv.Source
|
WHERE objectId=386942193651348;
|
+------+
|
| N |
|
+------+
|
| 39 |
|
+------+
|
1 row in set (0.09 sec)
|
|
|
mysql> select COUNT(*) AS N FROM qservTest_case01_qserv.Source
|
WHERE objectId IN(386942193651348, 386950783579546);
|
+------+
|
| N |
|
+------+
|
| 39 |
|
+------+
|
1 row in set (0.09 sec)
|
|
|
mysql> select COUNT(*) AS N FROM qservTest_case01_qserv.Source
|
WHERE objectId IN(386950783579546, 386942193651348);
|
+------+
|
| N |
|
+------+
|
| 56 |
|
+------+
|
1 row in set (0.11 sec)
|
Attachments
Issue Links
Activity
Hi Jacek Becla,
It's ok for me, thanks!
But, in the short term, I would like to finish my work on ORDER BY clause first. Are you OK with that? If yes, shall I rename DM-2859 to "Add support for ORDER BY clause" and implement here K.T. idea (i.e. adding a column in the select clause for each ORDER BY field and removing it after aggregation step?). I already invest a lot to implement this feature and I strongly think it should work and only require some simple code. Nevertheless FYI I think it should be at least 10SP.
Cheers
I tried to rebase my best but a few features are entangled. Hope it'll be easy to review. Interesting commit is here: 899dd2bf701971d10f3adb2fa17c43af542cc3e.
I needed to add tests/logging to chase this bug.
Please also review qserv_testdata.
Nice work Fabrice, I am done with the review, Andy, mark reviewed when done.
Fabrice, I'm done reviewing too. General comment - there is a lot of code refactoring beyond simply fixing broken IN. In fact I'm not really sure now what constitutes actual fix. I'd prefer all refactoring to go into separate ticket(s) (not just separate commits).
A bunch of comments and suggestions in PR. I do not close review yet, if you address all suggestions it will likely need another review 
Andy Salnikov, all comments have been adressed, thanks. Sorry for all the refactoring but I needed to improve logging/test code in order to chase the bug. And then it was easier to kept everything in the same ticket.
Fabrice, looks OK. I left few more minor comments on PR.
Fabrice, I think you will like this one. If not, give it back to me.