remdups - remove duplicate files
Project description
- Homepage:
- License:
See LICENSE file
remdups
remdups generates a script that removes duplicate files starting from file hashes. It can also generate a these file hashes. It allows to specify which files to keep and which to remove. Byte-wise comparison available to be on the safe side.
The resulting script should be further inspected, possibly re-generated with different parameters and finally executed separately in your shell.
Known Issues
- Linux:
None
- Windows:
A gave it a try on Windows Cmd, but ran into problems due to Python Issue 1602. So for file names with encodings falling into this issue it won’t work.
Install
Directly from PyPi:
$ pip install remdups
From github: Clone, change to directory and do
$ python setup.py install
If you plan to extend this tool
fork on github
clone locally from your fork
do an editable install
$ pip install -e .
test after change and bring coverage to 100%
$ py.test test_remdups.py --cov remdups.py --cov-report term-missing
consider sharing changes useful for others (github pull request).
Usage
The intended procedure to remove duplicates with remdups is:
create file hash list:
$ remdups --hash > hashes.txt
or
$ find . -not -type d -exec sha256sum {} \; > hashes.txt
make a script with remove commands
$ remdups [options] hashes.txt nodupes.sh
inspect the script and go back to 2., if necessary, else 4.
execute script
$ nodupes.sh
remove empty directories:
$ find . -empty -type d -delete
All in One
This takes long, because all the hashes are create anew. It is therefore not suitable to iterate with new parameters.
$ remdups
File Hash List
The file hash list as an intermediate starting point makes it faster to iterate with new parameters.
There are more ways to generate the file hash list.
Use find with a checksum generator
$ find . -not -type d -exec sha256sum {} \; > hashes.txt
Use remdups
remdups allows to make a file hash list with the --hash option and no input file.
$ remdups --hash > hashes.txt
With --hash one can use the --exclude-dir to ignore certain directories.
--hash together with a file can replace system checksum tools. remdups has these source options: --name, --namedate, --exif, --content, --block. For full content md5sum or shaXsum (X=1, 224, 256, 384, 512) system tools are faster.
Generate the remove script
You start with the file hash list
$ remdups [options] hashes.txt > rm.sh
or
$ remdups [options] hashes.txt rm.sh
At this stage you would use
-i and -o to choose which files get removed
-c to comment out the remove command
-r and -d to specify alternative remove commands for file and directory
-x to specify the extension used for html files subdirectory. It defaults to _files. If it starts with hyphen like -Dateien do -x="-Dateien".
-n --only-same-name to ignore duplicates with different name
-s --safe to do a final bytewise compare to make sure that files are really the same. You should add this option for the final remove script version. It can take a long time. After that you possibly still do manual changes to the script and then you execute it.
Help
Check out:
$ remdups –help
For use from within python check out the code.
Similar tools
I had to clean a sprawling directory and used python, then decided to make a little command line tool out of it.
Although I did a little googling before, only afterwards I found other links with similar tools. I did not test them because my directory is deduped by now.
I have made a Wikipedia page listing similar tools: List of Duplicate File Finders
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.