Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
0.9
-
None
-
None
Description
The following code works correctly when object reuse is switched off.
When switching it on, the results are wrong.
Using a string-based key selection (putting "name") works for both cases.
@Test public void testWithAvroGenericSer() throws Exception { final ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment(); // env.getConfig().enableObjectReuse(); Path in = new Path(inFile.getAbsoluteFile().toURI()); AvroInputFormat<User> users = new AvroInputFormat<User>(in, User.class); DataSet<User> usersDS = env.createInput(users); DataSet<Tuple2<String, Integer>> res = usersDS.groupBy(new KeySelector<User, String>() { @Override public String getKey(User value) throws Exception { return String.valueOf(value.getName()); } }).reduceGroup(new GroupReduceFunction<User, Tuple2<String, Integer>>() { @Override public void reduce(Iterable<User> values, Collector<Tuple2<String, Integer>> out) throws Exception { for(User u : values) { out.collect(new Tuple2<String, Integer>(u.getName().toString(), 1)); } } }); res.writeAsText(resultPath); res.print(); env.execute("Avro Key selection"); expected = "(Charlie,1)\n(Alyssa,1)\n"; }