aggregating tuples

[ktlina]

I have a workflow that executes different processes depending on the users preferences. When all desired processes are complete, a last process will take all the outputs and create a summary of the most relevant results. To do this, I need to take the output tuples generated by all processes and aggregate them into a single tuple. So for instance:

1. Process 1 output: ["sample1", /home/output/process1/sample1]
2. Process 2 output: ["sample1", /home/output/process2/sample1]
3. Process 3 output: ["sample1", /home/output/process3/sample1]

When using mix() and groupTuple() I am able to group the first two tuples and the result looks something like:
["sample1", [/home/output/process1/sample1, /home/output/process2/sample1]]. However, when I try to mix and grouTuple for the third process, I get this error:
'Not a valid path value type: nextflow.util.ArrayBag'
Mind you I'm trying to mix the resulting tuple (Tuple1/2) of mixing Tuple1 and Tuple2 with Tuple3.

I understand this is because I am trying to use an Array/List as the mix() method input, and this is not supported. So my question is, how can I group the process outputs in a tuple containing the sampleID and a list of paths? I need this in order to make the summary process as "relaxed" as possible, so that, even though the user decides not to run all processes, it will still work. This way I do not need to create an input for the summary process that looks like: tuple val(sampleID), path(process1), path(process2), path(process3) but instead, tuple val(sampleID), path(all_results).

I hope I have explained myself clearly.
Thnk you in advance

[bentsherman]

Can you share a minimal code example?

[ktlina]

Sorry, I just saw this. I kind of solved the problem by using "join" method and then mapping the tuples in the format I want them to be... so joining the tuple outputs of different processes (all_files = all_files.join(runMegaHitOutput, by: 0)) and then mapping them (final_tuple = all_files.map {tuple -> return [tuple[0], tuple[1..-1]]}) into a final tuple.

[bentsherman]

Glad you found a solution. join is probably what I would recommend for grouping related outputs from different processes, whereas groupTuple would be more for grouping related outputs from the same process. We are working to stamp out some of these quirks with the different operators so that it isn't such a hassle

[ktlina]

yeas, i feel like i have been having a lot ouf trouble when dealing with the combination of outputs from different processes, especially when working with multiple samples. thank you for you help, anyways!