Welcome to the Linux Foundation Forum!

Comparing contents of two directories using bash


I need help figuring out how I write a script that copies files that exist in one directory but don't exist in the second directory to the third directory.

I have the following scenario:

Files are uploaded from time to time to a directory. These are jpg, pdf, swf and a bunch of different files.

The files are then processed via script (pdfs are processed using imagemagick, jpgs copied to a certain directory and the names shortened and so forth) that is working very well. This is in part the reason for I need to change the workflow: The original file names cannot be compared to the output filenames. The whole process can take hours since we're talking about thousands of files.

The processing of the files (thousands) can take hours so I thought I'd change the workflow:

Create a script that would compare files from the "1-incoming" and "0-processed" directories and copy the files that haven't been processed (files that exist in "1-incoming" but not in "0-processed" to a directory called "2-toDo" but I am at a total loss how I would do that.

Here is an example:


The contents of which are:








With no files processed the "0-processed" directory is empty so what currently happens is that everything gets processed like this:

a) A script copies everything that exists in 1-incoming directory but doesn't exist in 0-processed to 2-toDo

b) A second script copies PDFs from 2-toDo to 3-toPDF and other files straight to 4-toTruncate

c) A third script uses imageMagick to convert PDFs to JPG files and copies the JPGs to 4-toTruncate and the PDFs to 0-processed

d) A fourth script copies the files in 4-toTruncate to 5-truncated under a shorter name and the original files to 0-processed

e) A fifth script copies files in 5-truncated to 6-toOutput

f) A sixth script copies the files in 6-toOutput to another server

I know that the last two steps can be joined in one step but I chose to write it up like this to make the example more clear.

I basically need help with step a)

Can anyone give me pointers as to how I would do this using bash?



  • blue_raptor
    " diff " and ' rsync ' commands should probably help you write a script for comparision and keeping sync between directories...
  • jabirali
    jabirali Posts: 157
    Since the filenames may not necessarily be the same after processing, you probably want to use md5sum to identify which files have already been processed. You could e.g. md5sum all the files in ./0-processed/ and store the hashes to a file like this:
    md5sum ./0-processed/* | sed 's/\s.*//' > hashes
    When deciding what files to process, you can then generate a list of files excluding any files matching the hash list like this:
    md5sum ./1-incoming/* | grep -vf hashes | sed 's/[0-9a-z]*\\s*//'
    For instance, you may loop over the files not matching the hash list and copy those files to ./2-toDo/ like this:
    for file in $( md5sum ./1-incoming/* | grep -vf hashes |
    sed 's/[0-9a-z]*\\s*//' ); do cp "$file" ./2-toDo/; done
    I hope this helps :)


Upcoming Training