Hey, thanks for replying so fast :)
I saw the discussion in a previous thread concerning changing the API to offer more explicit join functions. I think providing these special functions is a good way to disable any other kind of interaction with the solution set. However, as a user, I would like to have a more flexible way of interacting with the state of the iteration. In the program I'm describing above, I actually want to join on the value of the solution set, not the key. It would be nice to somehow have access to the solution set as any other normal DataSet. I'm not sure how this could be supported, but if you think this is a good idea, I could work on this! Cheers, V. On 31 July 2014 15:50, Stephan Ewen <se...@apache.org> wrote: > Hi Vasia! > > There is no fundamental reason, we simply have not gotten around to > implementing it, yet. Any help along these lines is highly welcome. > > One reason that held us back is that we need to make sure that the key of > the solution set and the key of the join is the same. > That is hard to verify with general functions. One approach is to actually > change the delta iteration API to define the keys only at > one place (the definition of the iteration), and offer special > "joinWithSolution" and "coGroupWithSolution" functions, rather then using > the regular join syntax (which allows to create invalid constructs). > > What are your thoughts on this, from a DeltaIteration user perspective? > > Greetings, > Stephan >
