Sort file names after using groupTuple

[zillurbmb51]

Hi, I have a similar issue #347 only have numbers in the file name.
The solution there:
my_channel.groupTuple().map { chrom,bgen -> tuple( chrom, bgen.sort{it.name})}
does not resolve the problem, because in my file name there are numbers existed. example:
[chr1, [chr1.step10.imputed.chunk0.bgen, chr1.step10.imputed.chunk1.bgen, chr1.step10.imputed.chunk2.bgen, chr1.step10.imputed.chunk3.bgen,.....]]
After using
.map { chrom,bgen -> tuple( chrom, bgen.sort{it.name})} it is giving me
[chr1, [chr1.step10.imputed.chunk0.bgen, chr1.step10.imputed.chunk1.bgen, chr1.step10.imputed.chunk10.bgen, chr1.step10.imputed.chunk11.bgen,.....]] . As I understand, sorting order was 0,1,10,11,12,... but I need, 0,1,2,3,4... That means

[chr1, [chr1.step10.imputed.chunk0.bgen, chr1.step10.imputed.chunk1.bgen, chr1.step10.imputed.chunk2.bgen, chr1.step10.imputed.chunk3.bgen,.....]] . How could I get that after using groupTuple ?

[abhi18av]

Hi @zillurbmb51 , for questions related to usage of Nextflow, please make use of the discussions forum. The issues sections are used for bugs/feature request regarding Nextflow runtime. This is a new functionality of Github and similar to issues, but dedicated to only questions.

[abhi18av]

Regarding the solution, I think that we could rely on some Groovy magic and closure over here

Given the following files in the filesystem

chr1.step10.imputed.chunk0.bgen
chr1.step10.imputed.chunk1.bgen
chr1.step10.imputed.chunk10.bgen
chr1.step10.imputed.chunk100.bgen
chr1.step10.imputed.chunk11.bgen
chr1.step10.imputed.chunk78.bgen

We want to soft them based on the NUM in ...chunkNUM.bgen pattern. For this we separate the NUM part and apply sort based on that.

Channel.fromPath("*.bgen")
       .set { fasta_ch }


fasta_ch
    .toSortedList { it -> (it.name =~ /chr1.step10.imputed.chunk(.*?).bgen/)[0][1].toInteger() }
    .flatten()
    .view()

This yields the desired result

/Users/eklavya/projects/code/_scratch/nextflow_playground/chr1.step10.imputed.chunk0.bgen
/Users/eklavya/projects/code/_scratch/nextflow_playground/chr1.step10.imputed.chunk1.bgen
/Users/eklavya/projects/code/_scratch/nextflow_playground/chr1.step10.imputed.chunk10.bgen
/Users/eklavya/projects/code/_scratch/nextflow_playground/chr1.step10.imputed.chunk11.bgen
/Users/eklavya/projects/code/_scratch/nextflow_playground/chr1.step10.imputed.chunk78.bgen
/Users/eklavya/projects/code/_scratch/nextflow_playground/chr1.step10.imputed.chunk100.bgen

[zillurbmb51]

Thanks a lot. After using the above solution, I am getting this:

My files are a part of tuples: [chr1, [chr1.step10.imputed.chunk0.bgen, chr1.step10.imputed.chunk1.bgen, chr1.step10.imputed.chunk2.bgen, chr1.step10.imputed.chunk3.bgen,.....]]
So, maybe the toSortedList operator need to be modified for tuples

These tuples were generated from:
imputed.groupTuple().transpose().map{chrom,bfiles -> tuple(chrom,bfiles[0])}.groupTuple()
Here, imputed is a output channel which produces multiple files for each chromosome. The groupTuple disrupted the ordering of files.
Need help here.

[abhi18av]

You'd need to customize the above solution I outlined, to fit your needs, for example.

 my_channel.groupTuple().map { chrom,bgen -> tuple( chrom, 
bgen.toSortedList { it -> (it.name =~ /chr1.step10.imputed.chunk(.*?).bgen/)[0][1].toInteger() })
}