[
https://issues.apache.org/jira/browse/TRAFODION-1562?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15652759#comment-15652759
]
ASF GitHub Bot commented on TRAFODION-1562:
-------------------------------------------
Github user zellerh commented on a diff in the pull request:
https://github.com/apache/incubator-trafodion/pull/826#discussion_r87315939
--- Diff: core/sql/optimizer/RelSequence.cpp ---
@@ -1179,20 +1179,21 @@ RelExpr *RelSequence::bindNode(BindWA *bindWA)
sequencedColumns() += newColumn->getValueId();
}
-
+ ValueIdMap ncToOldMap;
for(i = 0; i < colList->entries(); i++)
{
ValueId columnValueId = colList->at(i)->getValueId();
ItemExpr *newColumn = new (bindWA->wHeap())
NotCovered (columnValueId.getItemExpr());
newColumn->synthTypeAndValueId();
+ newColumn->setOrigOpType(columnValueId.getItemExpr()->origOpType());
resultTable->addColumn(bindWA,
colList->at(i)->getColRefNameObj(),
newColumn->getValueId(),
USER_COLUMN,
colList->at(i)->getHeading());
-
+ ncToOldMap.addMapEntry(columnValueId,newColumn->getValueId());
--- End diff --
These ValueIdMaps are usually oriented like the operators in the tree, with
values that are up the tree on top and those that are lower in the tree on the
bottom. In this case, the "not covered" are in the upper part of the tree and
the regular columns are lower. So, that's how they should go into the map, to
avoid confusion.
> Changes in query tree when the upsert command is transformed into merge
> -----------------------------------------------------------------------
>
> Key: TRAFODION-1562
> URL: https://issues.apache.org/jira/browse/TRAFODION-1562
> Project: Apache Trafodion
> Issue Type: Sub-task
> Components: sql-cmp
> Reporter: Selvaganesan Govindarajan
> Assignee: Sandhya Sundaresan
> Attachments: BatchUpsertTransformation.pdf
>
>
> to improve the performance as explained in the main JIRA
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
