I have multiple files (n=86000) with one column each and I want to combine them all into one file with 86000 columns.

I tried the following command

Paste *.txt > mean.txt

And as a for loop

for i in *.txt, do paste $i > mean.txt; done 

This did not appear to work because there are 86000 to combine, I then divided my files into multiple folders with each folder not containing more than 15000 files but this did not working

Format of my single file is as follows (It has 16000 lines)


What is the bug in the code?

  • 1
    $\begingroup$ Thanks. Do you want one long list? In this case the command is 'cat' not 'paste'. Just check you are not exceeding your disk space. $\endgroup$
    – M__
    Jul 10, 2019 at 10:56
  • 1
    $\begingroup$ What is the error message you get? $\endgroup$ Jul 10, 2019 at 11:08
  • 3
    $\begingroup$ Equivalently answered here assuming OP wants to concatenate vertically, not horizontally, though the approach could be modified anyway: stackoverflow.com/questions/18695105/… $\endgroup$
    – Joe Healey
    Jul 10, 2019 at 11:10
  • 3
    $\begingroup$ @DanielStandage such questions are on topic since manipulating and parsing text files is a very large part of a bioinformatician's work. This question is really not clear, mind you, but it is on topic. $\endgroup$
    – terdon
    Jul 10, 2019 at 16:23
  • 2
    $\begingroup$ @Priya tried to clarify the question a bit and I also suspected a typo in the code you tried. I hope the edits I have done are appropriate, if not please roll it back or edit it by your own :-). $\endgroup$ Jul 11, 2019 at 12:48

1 Answer 1


The bug in your loop appears be "> filename" per column, this over-rides the last iteration resulting in a file with a single column, which has been over-written 85999 times. Thus the file output you observe should be the single column present in the last file of your array. If you replaced "> filename" with ">> filename" within the loop you would then append to the file a new column with each iteration. If the loop fell over at some point you could then simply use one of the outputs of wc filename to assess how many columns you had (you might have to transponse the data for wc to work (lots of scripts for this in Perl), although there will almost certainly be an easy solution for counting columns in awk.

Personally, I would use a dataframe for this operation (possibly R, definately Python pandas), because 86000 columns is a lot of columns in any other walk of life, even if it is just post-processing. Generally, if you just want bash, awk is a better solution in this case (I don't code with it, but some here certainly do). However, I do get the idea that paste will work if each column is an identical length.

BTW paste must be lower case, you've got Paste in the first line of your code, but its okay in the second bit of code.

BTW BTW, I did not know shell would take *.txt directly as an array, with Perl/Python you need to "glob" it and assign it as an array/list.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.