Open jonasrk opened 7 years ago
Hi Jonas, first of all welcome to the dRHEEMing community!
Now, could you please give a bit more details about this issue? what do you mean with an API for the Loop operator? In other words, what would you expect Rheem to provide to make your life easier?
Hi Jorge, nice meeting you!
What I mean is this: The repeat
operator for example has its Java and Scala APIs defined in the package rheem-api
in the file DataQuanta.scala
in functions repeatJava
and repeat
.
The LoopOperator
does not have APIs like this defined (yet) and therefore can only be used in a more low-level fashion.
The API should look something like this:
val myLoopResult = planBuilder.load(...)
.loop { (convergence, data) => (convergence.map(...), data.filter(...).map(...) }
.collect()
Ok, It is clear now. I think this issue is a bug as each operator should be exposed properly to users. What do you think?
What's the difference with the doWhile(...)?
The DoWhileOperator
allows to change only a single dataset from iteration to iteration. The LoopOperator
, in contrast, permits to such mutable datasets.
This would make it much easier to work with.