python, unique files by content

I would like to retrieve a list of unique files by content rather than by filename.

That is, if spam.txt and eggs.txt both contained the same contents I want only one of them to return. A very simple approach is to compute a SHA-1 checksum on each file, and build a dictionary with the checksum as the unique key.

#!/usr/bin/env python
# vim: set tabstop=4 shiftwidth=4 autoindent smartindent:
import hashlib, sys
import logging

def _dupdoc(filelist):
	returns a list of unique files (by content rather than filename)
	that is, if spam.txt and eggs.txt both contained the same contents, 
	only one filename will be returned
	shasums = {}
	for file in filelist:
			fh = open(file, 'rb')
			sha1 = hashlib.sha1(
			if sha1 not in shasums:
				shasums[sha1] = file
				logging.debug('%s %s' %(file, sha1))
		except IOError as e:
			logging.warning('could not open %s' %(file))
	uniquelist = [file for file in shasums.values()]
	return uniquelist

if __name__ == "__main__":
	command-line, accept either a list of files in STDIN
	or a single filename argument that contains a list of files

	filelist = []
	if len(sys.argv) > 1:
		fh = open(sys.argv[1], 'r')
		filelist = fh.readlines()
		filelist = sys.stdin.readlines()
	filelist = [file.strip() for file in filelist]
	uniques = _dupdoc(filelist)
	for file in uniques:
		print file

The commandline __main__ portion of the program expects an optional command line argument, or if no argument is specified than a filelist will be read on STDIN, e.g.,

#  find test -type f | dupdoc
This entry was posted in python, shell tips. Bookmark the permalink.

Comments are closed.