Are there any `groupTuple`/`groupKey` magicians around? I’ve got a channel of

[meta, bam]

of which I want to gather and join the elements by

meta.samplename

. The catch is the channel elements can have a different amount of chuncks per

meta.samplename

. I’m looking for a non blocking way to count all the elements in channel per

meta.samplename

so I can make a

groupKey

Current code is in https://github.com/nf-cmgg/preprocessing/blob/dev/subworkflows/local/fastq_to_aligned_cram/main.nf#L68-L98. Here I first try to join on the amount of scattered

meta.chunks

and then on the amount of replicates in

meta.count

are you calling team sarek?

I invoke the ancient rites and call upon team #CGFUX04HZ to rid me of my woes

all hail

We're prepping a PR, but I'll ping the supreme sorceress of the groupKey magic once we're done

I figured out the specific issue in our pipeline but still interesting to see how you guys do it 👀

so what was the issue for you?

A meta map got the samplename ID instead of the lane ID before a groupTuple where we merge files from the same lane. This caused some weird behaviour when the lanes had different amount of files 😛

yeah, I understand the issue

We’re prepping a PR, but I’ll ping the supreme sorceress of the groupKey magic once we’re done

I’m still curious what you had in store though 😉