kgyrtkirk commented on a change in pull request #1912:
URL: https://github.com/apache/hive/pull/1912#discussion_r565902500
##########
File path:
ql/src/java/org/apache/hadoop/hive/ql/optimizer/SharedWorkOptimizer.java
##########
@@ -1791,9 +1792,16 @@ private static boolean validPreConditions(ParseContext
pctx, SharedWorkOptimizer
// If we do, we cannot merge. The reason is that Tez currently does
// not support parallel edges, i.e., multiple edges from same work x
// into same work y.
+ EdgePredicate edgePredicate;
+ if
(pctx.getConf().getBoolVar(ConfVars.HIVE_SHARED_WORK_PARALLEL_EDGE_SUPPORT)) {
+ edgePredicate = new EdgePredicate(EnumSet.<EdgeType> of(EdgeType.DPP,
EdgeType.SEMIJOIN, EdgeType.BROADCAST));
+ } else {
+ edgePredicate = new EdgePredicate(EnumSet.<EdgeType> of(EdgeType.DPP));
Review comment:
yes; the old behaviour was to not allow even those - however because DPP
edges are not physical tez edges they can be anywhere without restrictions; but
I could remove DPP as well to be extra safe
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
