[
https://issues.apache.org/jira/browse/CASSANDRA-9606?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14653664#comment-14653664
]
Sam Tunnicliffe commented on CASSANDRA-9606:
--------------------------------------------
Sorry this review has taken so long, it fell behind a couple of other tasks in
my backlog and then this is a particularly esoteric corner of the code (though
thankfully it does get clearer in later branches). I just have a few remarks:
In all three patches, there are changes to how partition key restrictions are
handled; explictly in the 2.0 & 2.1 patches in
{{SelectStatement#updateRestrictionsForRelation}} and implicitly (I think) in
2.2, but there aren't any new or updated tests that cover a change in
behaviour. Is it possible to add some?
In the 2.0 and 2.1 patches, {{SelectStatement#updateSingleColumnRestriction}}
had me scratching my head for a while, maybe a comment on the new check at line
1945/1953 might help clarify what's going on there.
Nits (can be fixed on commit)
In the 2.1 patch, there's an unused import in {{Restriction}} (it isn't new,
but may as well remove it)
In the 2.2 patch indentation error in {{MultiColumnRestriction.Slice}} where
the private modifier has been removed
The merge from 2.2 to trunk after applying the 2.2 patch generates a few
conflicts, but they're reasonably straightforward to resolve so I don't think
there's any need for a 3.0 specific patch.
> this query is not supported in new version
> ------------------------------------------
>
> Key: CASSANDRA-9606
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9606
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Environment: cassandra 2.1.6
> jdk 1.7.0_55
> Reporter: zhaoyan
> Assignee: Benjamin Lerer
> Attachments: 9606-2.0.txt, 9606-2.1.txt, 9606-2.2.txt
>
>
> Background:
> 1、create a table:
> {code}
> CREATE TABLE test (
> a int,
> b int,
> c int,
> d int,
> PRIMARY KEY (a, b, c)
> );
> {code}
> 2、query by a=1 and b<6
> {code}
> select * from test where a=1 and b<6;
> a | b | c | d
> ---+---+---+---
> 1 | 3 | 1 | 2
> 1 | 3 | 2 | 2
> 1 | 3 | 4 | 2
> 1 | 3 | 5 | 2
> 1 | 4 | 4 | 2
> 1 | 5 | 5 | 2
> (6 rows)
> {code}
> 3、query by page
> first page：
> {code}
> select * from test where a=1 and b<6 limit 2;
> a | b | c | d
> ---+---+---+---
> 1 | 3 | 1 | 2
> 1 | 3 | 2 | 2
> (2 rows)
> {code}
> second page：
> {code}
> select * from test where a=1 and b<6 and (b,c) > (3,2) limit 2;
> a | b | c | d
> ---+---+---+---
> 1 | 3 | 4 | 2
> 1 | 3 | 5 | 2
> (2 rows)
> {code}
> last page：
> {code}
> select * from test where a=1 and b<6 and (b,c) > (3,5) limit 2;
> a | b | c | d
> ---+---+---+---
> 1 | 4 | 4 | 2
> 1 | 5 | 5 | 2
> (2 rows)
> {code}
> question:
> this query by page is ok when cassandra 2.0.8.
> but is not supported in the latest version 2.1.6
> when execute：
> {code}
> select * from test where a=1 and b<6 and (b,c) > (3,2) limit 2;
> {code}
> get one error message：
> InvalidRequest: code=2200 [Invalid query] message="Column "b" cannot have
> both tuple-notation inequalities and single-column inequalities: (b, c) > (3,
> 2)"
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
