[
https://issues.apache.org/jira/browse/FLINK-1458?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14302966#comment-14302966
]
ASF GitHub Bot commented on FLINK-1458:
---------------------------------------
Github user StephanEwen commented on the pull request:
https://github.com/apache/flink/pull/357#issuecomment-72614433
Looks good to me so far, modulo two comments:
- Since the type extractor code is changed a lot, it would be good to
have a test validating that the TypeExtractor allows abstract types and
interfaces.
- I would change the log level of the messages (abstract types are less
performant, type is no POJO) to INFO, since WARN implies to me that something
is wrong (which is not)
> Interfaces and abstract classes are not valid types
> ---------------------------------------------------
>
> Key: FLINK-1458
> URL: https://issues.apache.org/jira/browse/FLINK-1458
> Project: Flink
> Issue Type: Bug
> Reporter: John Sandiford
> Assignee: Aljoscha Krettek
>
> I don't know whether this is by design or is a bug, but I am having trouble
> working with DataSet and traits in scala which is a major limitation. A
> simple example is shown below.
> Compile time warning is 'Type Main.SimpleTrait has no fields that are visible
> from Scala Type analysis. Falling back to Java Type Analysis...'
> Run time error is 'Interfaces and abstract classes are not valid types:
> interface Main$SimpleTrait'
> Regards, John
> val env = ExecutionEnvironment.getExecutionEnvironment
> trait SimpleTrait {
> def contains(x: String): Boolean
> }
> class SimpleClass extends SimpleTrait {
> def contains(x: String) = true
> }
> val data: DataSet[Double] = env.fromElements(1.0, 2.0, 3.0, 4.0)
> def f(data: DataSet[Double]): DataSet[SimpleTrait] = {
> data.mapPartition(iterator => {
> Iterator(new SimpleClass)
> })
> }
> val g = f(data)
> g.print()
> env.execute("Simple example")
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
