[
https://issues.apache.org/jira/browse/PHOENIX-2123?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14629367#comment-14629367
]
James Taylor commented on PHOENIX-2123:
---------------------------------------
Duplicate of PHOENIX-2110, [~samarthjain]. Are we ok not adding the column to
the PK of a diverged view? My thought on fixing this was that we'd need to
always push PK columns to even diverged views. KeyValue columns would not be
pushed. The cases that seem problematic are when the rows are operated on by
the table (post new PK column), then operated on (written) through the view
(which doesn't know about the PK columns). In this case, you wouldn't be able
to update the row as the unknown column values wouldn't be written.
> Diverged view prevents schema propagation to the entire view hierarchy
> ----------------------------------------------------------------------
>
> Key: PHOENIX-2123
> URL: https://issues.apache.org/jira/browse/PHOENIX-2123
> Project: Phoenix
> Issue Type: Test
> Reporter: Samarth Jain
> Assignee: Samarth Jain
> Fix For: 4.5.0
>
> Attachments: PHOENIX-2123.patch
>
>
> This code in MetadataEndPointImpl
> {code}
> if (view.getBaseColumnCount() ==
> QueryConstants.DIVERGED_VIEW_BASE_COLUMN_COUNT) {
> // if a view has divorced itself from the base table, we
> don't allow schema changes
> // to be propagated to it.
> return new
> MetaDataMutationResult(MutationCode.UNALLOWED_TABLE_MUTATION,
> EnvironmentEdgeManager.currentTimeMillis(), null);
> }
> {code}
> ends up preventing metadata changes to be propagated to the entire hierarchy.
> Because of this bug, some views might have the change and some may not which
> is not good.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
