A minimal example + query

[john-alexander]

Was trying to reproduce this example and came across a few issues with it

• p1.out.output_ch2 | p2 | p3 # there is no p3 process in the example
• Does both head.txt and tail.txt get stored in output_ch2?
  Needed to included gzip -f otherwise

Command error:
  gzip: tail.txt: Too many levels of symbolic links

Also I only seem to get tail.txt.gz in output_folder

Here's my modified version

#!/usr/bin/env nextflow
nextflow.enable.dsl = 2
params.outdir = "output_folder"
// Create input channel. Each .txt path is one element.
input_ch = Channel.fromPath( "*.txt" )  

workflow { 
 input_ch | p1
 p1.out.output_ch2 | p2 
}

process p1 {

 // Run locally instead of HPC or Cloud
 //executor 'local'  
 
 input:
   path(x) 

 output:
   path("head.txt")
   /* p2.out will include all output paths 
   from p2, whereas emit gives this 
   specific channel a name */
   path("tail.txt"), emit: output_ch2

 """
 head $x > head.txt
 tail $x > tail.txt
 """
}

process p2 {

 //executor 'local'
 // Copy paths out of the working directory
 publishDir params.outdir, mode:'copy'
 
 input:
   path(y)

 output:
   path("*.gz")

 """
 gzip $y
 """
}

[chlazaris]

Thank you, @john-alexander. I have now updated this. In order to save both .gzip files to the output directory without changing the process p2, I ended up saving the processes in modules, so that I call p2 twice in the workflow definition. There are probably other solutions too but this is the one that came to mind.

[chlazaris]

Now, this issue has been resolved