samples = Channel.fromPath(params.samplePath)
.splitCsv()
.subscribe{
row -> //How do I create a channel to run a process for each row
}
samples = Channel.fromPath(params.samplePath)
.splitCsv().into{channelOne; channelTwo}
maxErrorRetry
inside config file
withLabel: little_comp {
maxForks = 1
maxErrorRetry = 2
}
WARN: Unknown directive `maxErrorRetry` for process `msconvert`
I have a general question about design. I have a pipeline that looks like 'Y'; say processess 1->2->3->4 and then 4->5->7->9 and 4-6->8->10 are two possible continuations depending on a command line parameter, so either I take one direction, or the other direction, but in a single running instance it will always be the same. As I understand it, I need to create empty channels, e.g.
if (params.studyid > 0) {
ch_fastqs_dir = Channel.empty()
process irods {
...
Is there another way to go about this? If I need to create empty channels for all channels in the other branch, it becomes less readable. Perhaps some scoping could be introduced, where a set of processes are known to interact only with each other and channels are local to that scope? Ready to hide behind a :door:
when
declaration: https://www.nextflow.io/docs/latest/process.html#when
when:
, of course. Thanks!
when:
will be checked for each task. Is the overhead negligible, I wonder.