jdupes is a program based in fdupes. The main goal of jdupes is identify
and taking actions upon duplicate files. In comparison with fdupes, jdupes
is heavily modified from and improved.
.
The biggest reason to use jdupes is raw speed. In testing on various data
sets, jdupes is over 7 times faster than fdupes-1.51 on average.
.
Code in jdupes is written with data loss avoidance as the highest priority.
If a choice must be made between being aggressive or careful, the careful
way is always chosen.
.
jdupes includes features that are not found in fdupes. Examples of such
features include btrfs block-level deduplication and control over which
file is kept when a match set is automatically deleted. jdupes is not
afraid of dropping features of low value; a prime example is the -1 switch
which outputs all matches in a set on one line, a feature which was found
to be useless in real-world tests and therefore thrown out.
.
jdupes can convert duplicate files in hardlinks or relative softlinks. It
is useful in several scenarios, as in Debian packaging, to create relative
symlinks to lots of duplicate files (it will substitute rdfind + symlinks
commands, used to same purpose, when solving lintian duplicate-files).
.
Packages build for Linux versions have support to btrfs filesystem.
