Hello!
I'm new to Spark, and write a Java program with it.
But when I implement the Accumulator, something
confuse me, here is my implementation:
/class /WeightAccumulatorParam /implements /AccumulatorParam<Weight>{
/ @*Override*/
public Weight zero(/Weight /initialValue){
return new Weight();
}
/ @*Override*/
public Weight addInPlace(/Weight /v1, /Weight /v2){
return v1.add(v2);
}
}
just as the on-line documents says :
<http://apache-spark-user-list.1001560.n3.nabble.com/file/n16278/%E6%8D%95%E8%8E%B71.png>
link to the document
<https://spark.apache.org/docs/latest/programming-guide.html#parallelized-collections>
But then I met a mistake while building the file:
<http://apache-spark-user-list.1001560.n3.nabble.com/file/n16278/%E6%8D%95%E8%8E%B7.png>
It told me I need to override the function addAccumulator, too. But the
document doesn't say about this. So I am wonder whether I misunderstanding
the online document or the online document lost something?
And is it matter if I implement the addAccumulator just as addInPlace
function do?
--
View this message in context:
http://apache-spark-user-list.1001560.n3.nabble.com/Inconsistency-of-implementing-accumulator-in-Java-tp16278.html
Sent from the Apache Spark User List mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]