Repository: hive Updated Branches: refs/heads/master fafa953ff -> 99c0cafe5
HIVE-16947: Semijoin Reduction : Task cycle created due to multiple semijoins in conjunction with hashjoin (Deepak Jaiswal, reviewed by Jason Dere) Project: http://git-wip-us.apache.org/repos/asf/hive/repo Commit: http://git-wip-us.apache.org/repos/asf/hive/commit/99c0cafe Tree: http://git-wip-us.apache.org/repos/asf/hive/tree/99c0cafe Diff: http://git-wip-us.apache.org/repos/asf/hive/diff/99c0cafe Branch: refs/heads/master Commit: 99c0cafe5e40092fc4cf19074c9d5739b9ba6a49 Parents: fafa953 Author: Jason Dere <jd...@hortonworks.com> Authored: Wed Jun 28 10:58:57 2017 -0700 Committer: Jason Dere <jd...@hortonworks.com> Committed: Wed Jun 28 10:58:57 2017 -0700 ---------------------------------------------------------------------- .../hive/ql/optimizer/ConvertJoinMapJoin.java | 34 +++++++++++++------- 1 file changed, 23 insertions(+), 11 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/hive/blob/99c0cafe/ql/src/java/org/apache/hadoop/hive/ql/optimizer/ConvertJoinMapJoin.java ---------------------------------------------------------------------- diff --git a/ql/src/java/org/apache/hadoop/hive/ql/optimizer/ConvertJoinMapJoin.java b/ql/src/java/org/apache/hadoop/hive/ql/optimizer/ConvertJoinMapJoin.java index 98fec77..21d0053 100644 --- a/ql/src/java/org/apache/hadoop/hive/ql/optimizer/ConvertJoinMapJoin.java +++ b/ql/src/java/org/apache/hadoop/hive/ql/optimizer/ConvertJoinMapJoin.java @@ -888,19 +888,31 @@ public class ConvertJoinMapJoin implements NodeProcessor { } // Found a semijoin branch. - for (Operator<?> parent : mapjoinOp.getParentOperators()) { - if (!(parent instanceof ReduceSinkOperator)) { - continue; - } + // There can be more than one semijoin branch coming from the parent + // GBY Operator of the RS Operator. + Operator<?> parentGB = op.getParentOperators().get(0); + for (Operator<?> childRS : parentGB.getChildOperators()) { + // Get the RS and TS for this branch + rs = (ReduceSinkOperator) childRS; + ts = parseContext.getRsToSemiJoinBranchInfo().get(rs).getTsOp(); + assert ts != null; + for (Operator<?> parent : mapjoinOp.getParentOperators()) { + if (!(parent instanceof ReduceSinkOperator)) { + continue; + } - Set<TableScanOperator> tsOps = OperatorUtils.findOperatorsUpstream(parent, - TableScanOperator.class); - for (TableScanOperator parentTS : tsOps) { - // If the parent is same as the ts, then we have a cycle. - if (ts == parentTS) { - semiJoinMap.put(rs, ts); - break; + Set<TableScanOperator> tsOps = OperatorUtils.findOperatorsUpstream(parent, + TableScanOperator.class); + boolean found = false; + for (TableScanOperator parentTS : tsOps) { + // If the parent is same as the ts, then we have a cycle. + if (ts == parentTS) { + semiJoinMap.put(rs, ts); + found = true; + break; + } } + if (found) break; } } }
