Process fails to get file size when there is only 1 file in the input, but not when there is more than 1.

[ajmaurais]

Bug report

Expected behavior and actual behavior

In my minimal example I am trying to get the size of a collection of input files that is passed into a process. If there is 1 file in the input the GET_FILE_STATS process fails with No such file or directory but if there is more than 1 file the GET_FILE_STATS process will succeed.

Steps to reproduce the problem

main.nf

#!/usr/bin/env nextflow

nextflow.enable.dsl = 2

process GENERATE_FILE {
    input:
        val file_name

    output:
        path("$file_name")

    script:
    """
    echo "Hello there from ${file_name}" > '${file_name}'
    """
}

process GET_FILE_CONTENT {
    input:
        path input_files

    output:
        path('data.txt')

    script:
    """
    files=( '${input_files.join("' '")}' )

    echo "Hello from main" > data.txt

    for f in \${files[@]} ; do
        cat \$f >> data.txt
    done
    """
}

process GET_FILE_STATS {
    input:
        path input_files

    output:
        path('data.txt')

    script:
    """
    echo "There are ${input_files.size()} files" > data.txt
    echo "The total size is: ${input_files*.size().sum()} bytes" >> data.txt
    """
}

workflow {
    files_to_make = Channel.fromList(['file_1.txt'])
    // files_to_make = Channel.fromList(['file_1.txt', 'file_2.txt'])

    GENERATE_FILE(files_to_make)
    files = GENERATE_FILE.out

    GET_FILE_CONTENT(files.collect())
    file_data = GET_FILE_CONTENT.out.splitText()
    file_data.collect().view()

    GET_FILE_STATS(files.collect())
    file_data = GET_FILE_STATS.out.splitText()
    file_data.collect().view()
}

Program output

Failed execution when there is 1 file

In this example files_to_make = Channel.fromList(['file_1.txt'])

$ nextflow run main.nf

 N E X T F L O W   ~  version 25.04.4

Launching `main.nf` [adoring_cuvier] DSL2 - revision: 68eebc0ab7

executor >  local (2)
[79/5fc932] GENERATE_FILE (1) [100%] 1 of 1 ✔
[ef/4dc64b] GET_FILE_CONTENT  [100%] 1 of 1 ✔
[-        ] GET_FILE_STATS    -
['Hello from main\n', 'Hello there from file_1.txt\n']
ERROR ~ Error executing process > 'GET_FILE_STATS'

Caused by:
  No such file or directory: file_1.txt


Source block:
  """
  echo "There are ${input_files.size()} files" > data.txt
  echo "The total size is: ${input_files*.size().sum()} bytes" >> data.txt
  """

Tip: you can replicate the issue by changing to the process work dir and entering the command `bash .command.run`

 -- Check '.nextflow.log' file for details

Successful execution when there is more than 1 file

In this example files_to_make = Channel.fromList(['file_1.txt', 'file_2.txt'])

$ nextflow run main.nf

 N E X T F L O W   ~  version 25.04.4

Launching `main.nf` [shrivelled_linnaeus] DSL2 - revision: d96ebfaf04

executor >  local (4)
[39/f89949] GENERATE_FILE (1) [100%] 2 of 2 ✔
[c3/169d5d] GET_FILE_CONTENT  [100%] 1 of 1 ✔
[5e/d26150] GET_FILE_STATS    [100%] 1 of 1 ✔
['There are 2 files\n', 'The total size is: 56 bytes\n']
['Hello from main\n', 'Hello there from file_2.txt\n', 'Hello there from file_1.txt\n']

Environment

• Nextflow version: 25.04.4
• Java version: openjdk 21.0.7 2025-04-15
• Operating system: Ubuntu 24.04.2 LTS
• Bash version: GNU bash, version 5.2.21(1)-release (x86_64-pc-linux-gnu)

Additional context

I originally noticed this problem when I was trying to dynamically request memory in a process based off the size of the input files. If I were to add this: memory { Math.ceil(input_files*.size().sum() / (1024 ** 3))).GB } to the top of the GET_FILE_CONTENT process in the above example it would fail if there was only 1 file in input_files but not if there are more than 1. If I set the memory to a constant integer it will run successfully. The files are obviously getting transferred to the process work directory but there is something strange going on with file.size(). I have also tried using input_files.collect{ it.size() }.sum() but I get the same file not found error.

failing_nextflow.log