Anybody got an easy way to error out a pipeline if the samplesheet has fewer than N values? It's giving a run for my money... I've tried variations of this already, no dice. Gotten data type errors and errors that I believe are due to the asynchronous nature of the channels (also tried

.collect()

// Error if sample count is insufficient after filtering
if( ch_samplesheet.size() < params.min_samples ) {
    error "ERROR: At least ${params.min_samples} samples are required." }

Very common mistake. You cannot compare channels to static values like that. Also channels do not have a

size

but rather a

count

. Probably best to try something like:

ch_samplesheet.count().map{ ct -> ct < params.min_samples ? error("Too few samples. ${ct} < ${params.min_samples}") : "" }

But be aware that if you don’t block the use of

ch_samplesheet

downstream then processes can start before this full evaluates (they will then be killed/finish if this errors out). I can’t think of the best way to block in this instance right now, but I think there will be one.

If you are using nf-schema you can also add a check to the samplesheet schema: https://json-schema.org/understanding-json-schema/reference/array#length note that this does not show a nice error though so Arthur solution might be preffered

Ended up going with the nf-schema option since it was the cleanest, also the error message is clear enough. Albeit, I can't parameterize the minimum samples in my case but I don't need to in this case so it works

Awesome to hear that!