[
https://issues.apache.org/jira/browse/PHOENIX-385?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14375175#comment-14375175
]
James Taylor edited comment on PHOENIX-385 at 3/22/15 9:16 PM:
---------------------------------------------------------------
[~dhacker1341] - can you take a look at this one? It's likely the root cause of
the --1.00 issue you mentioned last week. See ParseNodeFactory.negate(ParseNode
child) and the hack to recognize expressions like -1*1. A reference to a
negative literal end up flowing this this code and turn it into -1*xxx. If we
recognize negative literals directly, we could remove that hack. The kind of
expressions that become problematic are SELECT 3-1, 3--1. I'm sure there's an
easy solution, but never had the time to figure it out.
was (Author: jamestaylor):
[~dhacker1341] - can you take a look at this one? It's likely the root cause of
the -1.00 issue you mentioned last week. See ParseNodeFactory.negate(ParseNode
child) and the hack to recognize expressions like -1*1. A reference to a
negative literal end up flowing this this code and turn it into -1*xxx. If we
recognize negative literals directly, we could remove that hack. The kind of
expressions that become problematic are SELECT 3-1, 3--1. I'm sure there's an
easy solution, but never had the time to figure it out.
> Support negative literal directly
> ---------------------------------
>
> Key: PHOENIX-385
> URL: https://issues.apache.org/jira/browse/PHOENIX-385
> Project: Phoenix
> Issue Type: Task
> Reporter: Raymond Liu
> Assignee: Dave Hacker
>
> say in split on / Limit etc who use literal as input
> Negative value should be supported, at present, only positive number is
> supported. Might not be true for Limit, but for other cases like split on
> might need it.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
