jdupes

Find duplicate files in a directory

$ jdupes [path/to/directory]

$ jdupes -r [path/to/directory]

$ jdupes -r -d [path/to/directory]

$ jdupes -r -dN [path/to/directory]

$ jdupes -r -S [path/to/directory]

$ jdupes -r -L [path/to/directory]

$ jdupes -r -m [path/to/directory]

$ jdupes -r [dir1] [dir2] [dir3]

jdupes [options] directory [directory ...]

jdupes is a fast duplicate file finder and remover. It identifies files with identical content by comparing file sizes, then partial hashes, then full hashes, and optionally byte-by-byte comparison for certainty.
The tool is a fork and enhancement of the older fdupes program, offering better performance (often 7x faster) and additional features. It uses optimized comparison algorithms and can handle millions of files efficiently.
Duplicates can be handled in several ways: interactive deletion (-d) presents choices for each set, automatic deletion (-dN) keeps the first file, and hard linking (-L) replaces duplicates with links to save space while preserving all filenames.
When scanning multiple directories, jdupes compares files across all of them, useful for finding duplicates between a backup and primary storage. The summary mode (-m) gives a quick overview of potential space savings.
The ordering option (-o) controls which file is considered "first" for automatic operations, allowing preference for keeping newer or alphabetically-first files.

-r, --recurse

Recurse into subdirectories.

Prompt for which duplicates to delete.

With -d, delete without prompting (keep first).

Show debug/diagnostic output.

Show size of duplicate files.

Print summary of duplicate statistics.

Replace duplicates with hard links.

Follow symbolic links.

Treat hard links as duplicates.

Exclude zero-length files.

Exclude hidden files.

Omit the first file in each duplicate set.

List each duplicate set on a single line.

Use null character as line terminator.

Hide progress indicator.

Skip byte-by-byte verification.

Exclude files by size criteria.

Order files by: name, time, none.

Hard linking only works on the same filesystem. Automatic deletion (-dN) with multiple directories may not preserve the expected file. Very large file counts may require significant memory. Quick mode (-Q) skips verification and could theoretically produce false positives.

jdupes was created by Jody Bruchon starting around 2015 as an enhanced fork of fdupes. The primary goals were improved performance and additional features while maintaining compatibility. The name combines "Jody's" with "dupes." It has become a popular choice for duplicate management, available in most Linux distribution repositories.

fdupes(1), rdfind(1), duperemove(8), rmlint(1)