Github user JamesRTaylor commented on a diff in the pull request:
https://github.com/apache/phoenix/pull/303#discussion_r200207759
--- Diff:
phoenix-core/src/main/java/org/apache/phoenix/schema/MetaDataClient.java ---
@@ -1957,6 +1968,17 @@ private PTable
createTableInternal(CreateTableStatement statement, byte[][] spli
linkStatement.setLong(6, parent.getSequenceNumber());
linkStatement.setString(7,
PTableType.INDEX.getSerializedValue());
linkStatement.execute();
+
+ // Add row linking index table to parent table for indexes
on views
+ if (parent.getType() == PTableType.VIEW) {
+ linkStatement =
connection.prepareStatement(CREATE_VIEW_INDEX_PARENT_LINK);
+ linkStatement.setString(1, tenantIdStr);
+ linkStatement.setString(2, schemaName);
+ linkStatement.setString(3, tableName);
+ linkStatement.setString(4,
parent.getName().getString());
+ linkStatement.setByte(5,
LinkType.VIEW_INDEX_PARENT_TABLE.getSerializedValue());
+ linkStatement.execute();
+ }
--- End diff --
We need to update MetaDataClient.createTableInternal() to not include the
columns from the parent table in 4.15 so that we can remove the code in
MetaDataEndPointImpl that filters the columns. It's fine to do this in a follow
up JIRA, but we should remember to do it.
---
