[
https://issues.apache.org/jira/browse/PHOENIX-1002?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14054632#comment-14054632
]
James Taylor commented on PHOENIX-1002:
---------------------------------------
Thanks for the patch, [~kbuzsaki]. Same comment as with PHOENIX-1055 wrt going
through codec. Also, would you mind filing a separate JIRA for this (i.e.
support for a MOD built-in)? I think this is a good short term solution for
[~tdsilva].
Would you also be up for tackling support the mod operator? You start with
modifying the grammar (PhoenixSQL.g), adding the % operator based on the
precedence (I think it's the same precedence as * and /)? Then you'd add a new
ParseNode class followed by new visitor methods in ParseNodeVisitor. Then you
implement them following the pattern of the others - the important one to
implement is in ExpressionCompiler. You can likely reuse/copy/paste from the
other implementations. It's a bit simpler than the other ones, b/c you don't
have to support all the various types (i.e. double, float, decimal), but can
just use a long. Might be able to just create a new ModulusFunction from
ExpressionCompiler to reuse your built-in.
> Add support for % operator
> --------------------------
>
> Key: PHOENIX-1002
> URL: https://issues.apache.org/jira/browse/PHOENIX-1002
> Project: Phoenix
> Issue Type: New Feature
> Reporter: Thomas D'Silva
> Attachments: PHOENIX-1002.patch
>
>
> Supporting the % operator would allow using sequences to generate IDs that
> are less than a maximum number.
> CREATE SEQUENCE foo.bar
> SELECT ((NEXT VALUE FOR foo.bar)%1000)
--
This message was sent by Atlassian JIRA
(v6.2#6252)
