[
https://issues.apache.org/jira/browse/PHOENIX-1271?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14144496#comment-14144496
]
Karel Vervaeke commented on PHOENIX-1271:
-----------------------------------------
My fugly knee-jerk solution was something along these lines
+ if (connection.getTenantId() != null) {
+ // exclude the first PK for all multi tenant tables
+ for (PTable table:
Iterators.toArray(connection.getMetaDataCache().iterator(), PTable.class)) {
+ if (!table.isMultiTenant()) {
+ continue;
+ }
+ appendConjunction(where);
+ // this is just a test, ignoring the conditions on schema and
catalog for now
+ where.append(String.format("(TABLE_NAME <> '%s' OR
ORDINAL_POSITION > 2)", table.getName()));
+
+ }
+ }
Not exactly elegant + multi-tenant tables now start with ORDINAL_POSITION=2.
I'll try James' first suggestion and see there what takes me.
> Column metadata doesn't hide tenant column on tenant-specific connections.
> --------------------------------------------------------------------------
>
> Key: PHOENIX-1271
> URL: https://issues.apache.org/jira/browse/PHOENIX-1271
> Project: Phoenix
> Issue Type: Bug
> Reporter: Karel Vervaeke
>
> connection.getMetaData().getColumns(...) always returns all columns.
> For multi-tenant tables it would make more sense if the first PK (the
> tenant-id column) would be excluded. If we do this the number of columns
> returned would match the number of columns you get when executing a "select
> *" query.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
