----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/26435/#review55851 -----------------------------------------------------------
data/conf/hive-log4j.properties <https://reviews.apache.org/r/26435/#comment96225> are ObjectStore and Operator lines intended? they'd affect logging that might be useful for other tests ql/src/java/org/apache/hadoop/hive/ql/exec/MapOperator.java <https://reviews.apache.org/r/26435/#comment96232> is it possible to document those with a small comment; as well as MapOpMeta and MapOpCtx classes? ql/src/java/org/apache/hadoop/hive/ql/exec/MapOperator.java <https://reviews.apache.org/r/26435/#comment96226> nit: hash codes could be combined similar to how Java does it ( * prime + next) ql/src/java/org/apache/hadoop/hive/ql/exec/MapOperator.java <https://reviews.apache.org/r/26435/#comment96227> why are the calles to SerDeUtils gone? Just asking ql/src/java/org/apache/hadoop/hive/ql/exec/MapOperator.java <https://reviews.apache.org/r/26435/#comment96228> this creates a context and checks for its presence, but then puts the result of some call taking the context to map. Would it make sense to put context into map first and then do processing on it (that I assume the call does)? That would be less confusing. ql/src/java/org/apache/hadoop/hive/ql/exec/MapOperator.java <https://reviews.apache.org/r/26435/#comment96229> is children.size() check no longer necessary? ql/src/java/org/apache/hadoop/hive/ql/exec/MapOperator.java <https://reviews.apache.org/r/26435/#comment96230> path is never used ql/src/java/org/apache/hadoop/hive/ql/exec/MapOperator.java <https://reviews.apache.org/r/26435/#comment96233> one child operator can be present for several paths or in several MapOpCtx? - Sergey Shelukhin On Oct. 8, 2014, 4:06 a.m., Navis Ryu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/26435/ > ----------------------------------------------------------- > > (Updated Oct. 8, 2014, 4:06 a.m.) > > > Review request for hive. > > > Bugs: HIVE-8186 > https://issues.apache.org/jira/browse/HIVE-8186 > > > Repository: hive-git > > > Description > ------- > > {noformat}select t1.BLOCK__OFFSET__INSIDE__FILE,t2.BLOCK__OFFSET__INSIDE__FILE > from src t1 join src t2 on t1.key = t2.key;{noformat} > Passes > {noformat}select t2.BLOCK__OFFSET__INSIDE__FILE > from src t1 join src t2 on t1.key = t2.key;{noformat} > Fails. > > The issue is that LazyBinarySerDe OI receives data intended for > UnionStructObjectInspector. > Judging by the above it has something to do with scanning table once for two > aliases. > > I'll look tomorrow > > > Diffs > ----- > > data/conf/hive-log4j.properties 7f5dfc4 > ql/src/java/org/apache/hadoop/hive/ql/exec/MapOperator.java f624bf4 > ql/src/java/org/apache/hadoop/hive/ql/exec/Operator.java 3dc7c76 > ql/src/java/org/apache/hadoop/hive/ql/exec/ReduceSinkOperator.java d8698da > ql/src/java/org/apache/hadoop/hive/ql/exec/Utilities.java 155002a > ql/src/java/org/apache/hadoop/hive/ql/exec/vector/VectorMapOperator.java > 311f6d6 > ql/src/java/org/apache/hadoop/hive/ql/plan/TableDesc.java 78d4d1f > ql/src/test/org/apache/hadoop/hive/ql/exec/TestOperators.java 90e4cad > ql/src/test/queries/clientpositive/join_vc.q 63b3da7 > ql/src/test/results/clientpositive/join_vc.q.out 12004ca > > Diff: https://reviews.apache.org/r/26435/diff/ > > > Testing > ------- > > > Thanks, > > Navis Ryu > >
