I have two main directories say, A and B. Each of them contains 100 sub directories (a1,a2,.. and b1,b2,...) with random name. Those sub directories contain a file of XY type. The X value is same for a particular sub directory say a1 and b1 but the Y value is different. I now can create a XY type file where Y will be the difference between the sub directories Y column only, say Y = Y1(a1) - Y2(b1) with a simple awk command or so. But I don't know how to do this with a bash loop for all the sub directories. What I want a desired script which will create directories 1, 2,.. where each of the directory will contain a XY type file and the XY type files contain X Y(a1-b1) in directory 1, X Y(a2-b2) in directory 2,.., X Y(a100-b100) in direcotry 100. I am quite new in bash so apologize for any trivial question.

Schematic example can be like this: I have two main directories A and B. A has sub directories 0.3427/ 0.3514/ 0.3543/ 0.3792/... and B has sub directories 0.4011/ 0.4014/ 0.4031/ 0.4357/....

0.3427/ sub directory of A contains the data file:

0.00 -8.79583
50.00 -8.79621
100.00 -8.79961
150.00 -8.80721
200.00 -8.81845
250.00 -8.83271
300.00 -8.8495

and 0.4011 sub directory contains the data file:

0.00 -8.78888
50.00 -8.78954
100.00 -8.79458
150.00 -8.80459
200.00 -8.81852
250.00 -8.83561
300.00 -8.85529

Now I want in directory 1 there should be a file say data.dat with the following data:

0.00 0.00695
50.00 0.00667
100.00 0.00503
150.00 0.00262
200.00 -7e-05
250.00 -0.0029
300.00 -0.00579

This desired data.dat file is just the difference of column Y of first sub directories of A and B which can be produce by an awk/paste command. Now I want the same for all the other sub directories present in the A and B. where in directory 2 the data.dat file should contain the difference in Y column of 0.3514/ of A and 0.4014/ of B and so on.

  • For a question like this it's sometimes easier to provide examples; what you have before and what you have afterwards. All the words you've written are unclear on the actual task. Aug 23, 2016 at 14:30
  • @Stephen: Al right, I will edit the question with examples shortly.
    – baban
    Aug 23, 2016 at 14:32
  • @Stephen: I tried to make the question a bit clear with example. Does it make sense now? sorry, for not writing properly.
    – baban
    Aug 23, 2016 at 14:48

1 Answer 1


Try this one.


# Run it in the folder containing dir A and dir B.

A=A #first dir, replace with the propper name
B=B #second dir, replace with the propper name
in_file=f #replace with propper name

B_subs=( $(ls B | sort -n) )

cd "$A"
for d in $( ls | sort -n) ; do
    mkdir ../"$i"
    # the one below is not correct, but gives your results
    # join "$a" "$b" | awk '{print $1, $3-$2}' > ../"$i"/"$out_file"
    # the one below is correct
    join "$a" "$b" | awk '{print $1, $2-$3}' > ../"$i"/"$out_file"
  • Thank you so much, this is perfect answer, code is working perfectly and giving me desire outputs. Sorry, I can't upvote your answer which made my life easy.
    – baban
    Aug 23, 2016 at 18:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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