[ 
https://issues.apache.org/jira/browse/CALCITE-2102?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16299673#comment-16299673
 ] 

Julian Hyde commented on CALCITE-2102:
--------------------------------------

A few comments:
* I backed out your changes to the parser ({{SqlOrderBy}} and 
{{SqlParserTest}}). The parser should not attempt to "optimize".
* In my branch 
https://github.com/julianhyde/calcite/tree/2102-duplicate-sort-keys I made 
RelCollations.of eliminate RelFieldCollations of a field seen ordinal.
* I don't understand the purpose of your change to 
{{RelCollations.contains(List<RelCollation> collations, ImmutableIntList 
keys)}}. Could you add a unit test - similar to {{RelCollationTest.contains}} - 
to help me understand.


> SqlToRelConverter (or perhaps RelBuilder) ignore repeat sort keys. In 
> addition, it should also prevent the incompatible sort keys.
> ----------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: CALCITE-2102
>                 URL: https://issues.apache.org/jira/browse/CALCITE-2102
>             Project: Calcite
>          Issue Type: Bug
>            Reporter: John Fang
>            Assignee: Julian Hyde
>
>         I test the  query {{SELECT a, c FROM MyTable ORDER BY a DESC, a 
> DESC}}. And the collation of LogicalSort is {{[0 DESC, 0 DESC]}} , thus it 
> will order by the repeater fields. I think the calcite should removal the 
> repeated fields, thus it make the LogicalSort whose collation is {{[0 DESC]}} 
> be more reasonable. 
>       Considering compatibility,it may be better make {{SqlToRelConverter}} 
> (or perhaps RelBuilder) ignore sort keys that have appeared earlier in the 
> key. In addition, it should also prevent the incompatible sort keys. It 
> should throw illegalargumentexception for the query {{SELECT a, c FROM 
> MyTable ORDER BY a DESC, a ASC}} because of the incompatibility.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Reply via email to