finddupes

Finds duplicate files based on hash and deletes them based on a given pattern.

Background

finddupes tries to be efficient by

• comparing file size before running expensive hash calculations
• using hash tables to find duplicate sizes/hashes in constant time on avg
• using the fast xxHash algorithm to calculate hashes
• running things in parallel. However, this only really helps if directories to be searched for reside on different media
• using an optional "cache" that can be reused and extended for multiple searches/deletions

What does finddupes not do

• try to find very similar files (fuzzy search)

Usage

Run finddupes with the -help flag to get all options:

finddupes -help

The execution can be interrupted with Ctrl-c. This will gracefully finish all calculation and write operations before shutting down.

Find duplicates in given directories

This will list all found duplicates.

finddupes <path> [path...]

Depending on the amount and size of files this can take a long time. For a large amount of files it is recommended to index all duplicates and store them in a database file. See next section.

Index files

Index all files from given directories recursively and store them in a database file.

finddupes -verbose -storeonly -path <db file path> <path> [path...]

e.g.

finddupes -verbose -storeonly -path pics.db ~/Pictures ~/Videos ~/DCIM

After indexing files one or more actions can be run to delete duplicates. A single last file will be always kept, regardless if there's a match or not.

The default is a dry run. To actually delete files, add the -delete flag.

Alternatively to indexing first, all actions can be run on the fly by not passing the -path <db file path> parameter.

finddupes -delmatch <pattern> ~/Pictures ~/Videos

See the next sections for a list of possible actions.

Delete duplicates based on a pattern

Delete duplicates whose path matches the given regex.

finddupes -path <db file path> -delmatch <pattern>

e.g.

finddupes -path pics.db -delmatch '\.jpe?g$'

Keep duplicates based on a pattern

Keep duplicates whose path matches the given regex.

finddupes -path <db file path> -keepmatch <pattern>

e.g.

finddupes -path pics.db -keepmatch '_orignal$'

Keep most recent duplicate

Keep the most recent duplicate, delete all others. Based on modification time (mtime).

finddupes -path <db file path> -keeprecent

Keep oldest duplicate

Keep the oldest duplicate, delete all others. Based on modification time (mtime).

finddupes -path <db file path> -keepoldest

Keep first duplicate

Keep the first duplicate based on lexically sorted file paths (not file names), delete all others.

finddupes -path <db file path> -keepfirst

Keep last duplicate

Keep the last duplicate based on lexically sorted file paths (not file names), delete all others.

finddupes -path <db file path> -keeplast