[
https://issues.apache.org/jira/browse/PHOENIX-4345?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16783832#comment-16783832
]
Geoffrey Jacoby commented on PHOENIX-4345:
------------------------------------------
[~yanxinyi] - Thanks for the changes. I thought about this some more. As a
result of this change, will the client developer get an IndexNotFoundException
back if they submit a statement referencing a non-existent index?
This is a change in behavior, since before clients just had to catch
TableNotFoundException if they wanted to have special error handling logic for
incorrect table and index names. (I know, for example, that we do this in some
of our internal tools.)
Wondering if it would be better to have IndexNotFoundException inherit from
TableNotFoundException so that existing catch blocks in client code don't need
to be revised. What do you think?
> Error message for incorrect index is not accurate
> -------------------------------------------------
>
> Key: PHOENIX-4345
> URL: https://issues.apache.org/jira/browse/PHOENIX-4345
> Project: Phoenix
> Issue Type: Bug
> Reporter: Ethan Wang
> Assignee: Xinyi Yan
> Priority: Trivial
> Attachments: PHOENIX-4345-master.patch, PHOENIX-4345.patch,
> PHOENIX-4345.patch, PHOENIX-4345v2.patch
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Error message for incorrect index is not accurate. it shows "table
> undefined". rather, should be index undefined.
> Table name: PERSON
> Index name: LOCAL_ADDRESS
> 0: jdbc:phoenix:localhost:2181:/hbase> ALTER INDEX LOCAL_ADDRESSX ON PERSON
> rebuild;
> Error: ERROR 1012 (42M03): Table undefined. tableName=LOCAL_ADDRESSX
> (state=42M03,code=1012)
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
