[
https://issues.apache.org/jira/browse/CALCITE-624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14365846#comment-14365846
]
Jesus Camacho Rodriguez commented on CALCITE-624:
-------------------------------------------------
[~julianhyde], [~mmokhtar] and myself hit the same problem today; it appears
sporadically.
Yesterday we talked a bit about the Comparator used in the TreeSet, and that
maybe it could be improved. Could you elaborate on that?
CC'd [~jpullokkaran]
> IllegalArgumentException: duplicate key in ReflectiveRelMetadataProvider
> ------------------------------------------------------------------------
>
> Key: CALCITE-624
> URL: https://issues.apache.org/jira/browse/CALCITE-624
> Project: Calcite
> Issue Type: Bug
> Reporter: Jesus Camacho Rodriguez
>
> We have our own implementation of RelMDParallelism provider in Hive:
> {{HiveRelMDParallelism}}. It has two methods for the HiveJoin class:
> {{isPhaseTransition}} and {{splitCount}}. We hit the problem shown in the
> stacktrace, that we have not been able to reproduce again.
> {noformat}
> java.lang.IllegalArgumentException: duplicate key: class
> org.apache.hadoop.hive.ql.optimizer.calcite.reloperators.HiveJoin
> at
> com.google.common.base.Preconditions.checkArgument(Preconditions.java:119)
> at
> com.google.common.collect.RegularImmutableMap.<init>(RegularImmutableMap.java:67)
> at
> com.google.common.collect.ImmutableMap.copyOf(ImmutableMap.java:303)
> at
> org.apache.calcite.rel.metadata.ReflectiveRelMetadataProvider.reflectiveSource(ReflectiveRelMetadataProvider.java:204)
> at
> org.apache.calcite.rel.metadata.ReflectiveRelMetadataProvider.reflectiveSource(ReflectiveRelMetadataProvider.java:112)
> at
> org.apache.hadoop.hive.ql.optimizer.calcite.stats.HiveRelMdParallelism.getMetadataProvider(HiveRelMdParallelism.java:44)
> at
> org.apache.hadoop.hive.ql.optimizer.calcite.HiveDefaultRelMetadataProvider.getMetadataProvider(HiveDefaultRelMetadataProvider.java:52)
> at
> org.apache.hadoop.hive.ql.parse.CalcitePlanner$CalcitePlannerAction.apply(CalcitePlanner.java:760)
> at
> org.apache.hadoop.hive.ql.parse.CalcitePlanner$CalcitePlannerAction.apply(CalcitePlanner.java:713)
> at org.apache.calcite.tools.Frameworks$1.apply(Frameworks.java:109)
> at
> org.apache.calcite.prepare.CalcitePrepareImpl.perform(CalcitePrepareImpl.java:730)
> at
> org.apache.calcite.tools.Frameworks.withPrepare(Frameworks.java:145)
> at
> org.apache.calcite.tools.Frameworks.withPlanner(Frameworks.java:105)
> at
> org.apache.hadoop.hive.ql.parse.CalcitePlanner.getOptimizedAST(CalcitePlanner.java:580)
> ...
> {noformat}
> We realized that the same class was ending up twice in a Set when the methods
> are added for reflection calls.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
