Re: problem with union

Posted by Michele Bertoni on
URL: http://deprecated-apache-flink-user-mailing-list-archive.369.s1.nabble.com/problem-with-union-tp2043p2542.html

Hi everybody,

I am still waiting for version 0.9.1 to solve this problem, any idea on when it will be released?

Thanks

Best,

michele

Il giorno 15/lug/2015, alle ore 15:58, Maximilian Michels <[hidden email]> ha scritto:

I was able to reproduce this problem. It turns out, this has already been fixed in the snapshot version: https://issues.apache.org/jira/browse/FLINK-2229

The fix will be included in the upcoming 0.9.1 release. Thank you again for reporting!

Kind regards,

Max

On Wed, Jul 15, 2015 at 11:33 AM, Maximilian Michels <[hidden email]> wrote:

Hi Michele,

Thanks for reporting the problem. It seems like we changed the way we compare generic types like your GValue type. I'm debugging that now. We can get a fix in for the 0.9.1 release.

Cheers,

Max

On Tue, Jul 14, 2015 at 5:35 PM, Michele Bertoni <[hidden email]> wrote:

Hi everybody, this discussion started in an other thread about a problem in union, but you said it was a different error then i am opening a new topic

I am doing the union of two dataset and I am getting this error




Exception in thread "main" org.apache.flink.api.common.InvalidProgramException: Cannot union inputs of different types. Input1=scala.Tuple6(_1: Long, _2: String, _3: Long, _4: Long, _5: Character, _6: ObjectArrayTypeInfo<GenericType<it.polimi.genomics.core.DataTypes.GValue>>), input2=scala.Tuple6(_1: Long, _2: String, _3: Long, _4: Long, _5: Character, _6: ObjectArrayTypeInfo<GenericType<it.polimi.genomics.core.DataTypes.GValue>>)
        at org.apache.flink.api.java.operators.UnionOperator.<init>(UnionOperator.java:46)
        at org.apache.flink.api.scala.DataSet.union(DataSet.scala:1101)
        at it.polimi.genomics.flink.FlinkImplementation.operator.region.GenometricCover2$.apply(GenometricCover2.scala:125)
        ...




Input1=
scala.Tuple6(_1: Long, _2: String, _3: Long, _4: Long, _5: Character, _6: ObjectArrayTypeInfo<GenericType<it.polimi.genomics.core.DataTypes.GValue>>)
input2=
scala.Tuple6(_1: Long, _2: String, _3: Long, _4: Long, _5: Character, _6: ObjectArrayTypeInfo<GenericType<it.polimi.genomics.core.DataTypes.GValue>>)


as you can see the two datasets have the same type
this error only happens with a custom data type (e.g. i am using an array of GValue, an array of Int or Double works)

in the last flink version it was working (milestone and snapshot) now in 0.9.0 it is not

what can it be?


thanks for help

cheers,
Michele