[
https://issues.apache.org/jira/browse/CASSANDRA-9970?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14697609#comment-14697609
]
Benjamin Lerer commented on CASSANDRA-9970:
-------------------------------------------
The new unit tests caught a problem in the 3.0 branch. I found that there was
an issue with the descending ordering for IN relations on the partition key
columns.
I added some extra tests.
* The new patch for 2.2 is
[here|https://github.com/apache/cassandra/compare/trunk...blerer:9970]
* The patch for 3.0 is
[here|https://github.com/apache/cassandra/compare/trunk...blerer:9970-3.0]
The results of the unit tests are
[here|http://cassci.datastax.com/view/Dev/view/blerer/job/blerer-9970-testall/lastCompletedBuild/testReport/]
and
[here|http://cassci.datastax.com/view/Dev/view/blerer/job/blerer-9970-3.0-testall/lastCompletedBuild/testReport/]
The results of the dtests are
[here|http://cassci.datastax.com/view/Dev/view/blerer/job/blerer-9970-dtest/lastCompletedBuild/testReport/]
and
[here|http://cassci.datastax.com/view/Dev/view/blerer/job/blerer-9970-3.0-dtest/lastCompletedBuild/testReport/]
> CQL 3.3 ordering in where clause produces different results when using
> "IN...ORDER BY"
> --------------------------------------------------------------------------------------
>
> Key: CASSANDRA-9970
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9970
> Project: Cassandra
> Issue Type: Bug
> Components: API
> Reporter: Marc Zbyszynski
> Assignee: Benjamin Lerer
> Priority: Minor
> Fix For: 2.2.x
>
> Attachments: 9970.txt, dtest-CASSANDRA-9970.txt
>
>
> This is an issue I found with
> {code}
> [cqlsh 5.0.1 | Cassandra 2.2.0 | CQL spec 3.3.0 | Native protocol v4]
> {code}
> Given this table and data:
> {code:sql}
> cqlsh> CREATE TABLE myschema.in_test (
> ... col_1 int,
> ... col_2 int,
> ... col_3 int,
> ... PRIMARY KEY ((col_1, col_2), col_3)
> ... );
> cqlsh> INSERT INTO myschema.in_test (col_1, col_2, col_3) VALUES(1, 1, 1);
> cqlsh> INSERT INTO myschema.in_test (col_1, col_2, col_3) VALUES(1, 1, 2);
> cqlsh> INSERT INTO myschema.in_test (col_1, col_2, col_3) VALUES(1, 1, 13);
> cqlsh> INSERT INTO myschema.in_test (col_1, col_2, col_3) VALUES(1, 2, 10);
> cqlsh> INSERT INTO myschema.in_test (col_1, col_2, col_3) VALUES(1, 2, 11);
> {code}
> This query behaves as expected (as described in
> [http://docs.datastax.com/en/cql/3.3/cql/cql_using/useQueryIN.html])
> {code:sql}
> cqlsh> select * from myschema.in_test where col_1=1 and col_2 IN (1,2) order
> by col_3 desc;
> InvalidRequest: code=2200 [Invalid query] message="Cannot page queries with
> both ORDER BY and a IN restriction on the partition key; you must either
> remove the ORDER BY or the IN and sort client side, or disable paging for
> this query"
> {code}
> But if you swap the order of the statements in the where clause, the query
> executes without error and doesn't totally honor the _order by_ clause:
> {code:sql}
> cqlsh> select * from myschema.in_test where col_2 IN (1,2) and col_1=1 order
> by col_3 desc;
> col_1 | col_2 | col_3
> -------+-------+-------
> 1 | 2 | 11
> 1 | 2 | 10
> 1 | 1 | 13
> 1 | 1 | 2
> 1 | 1 | 1
> (5 rows)
> {code}
> Apologies in advance if this is working as intended, or I'm duplicating an
> existing issue. Also please let me know if you need any additional info.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
