Skip to main content

Renames files using regular expressions

Project description

Renames files using regular expression matching. This enables elegant handling of multiple renames using a single command.


Basic syntax:

rename [-I] [-l] [-q] [-t] [-u] [-v "except_regex"] "regex" "target"

rename [--selftest]



Regular expression that matches source files which are to be renamed. Examples:


The regular expression is global by default (e.g. writing "[0-9]" means "^[0-9]$"). This is to avoid accidental partial catches. If you want to match all files that start or end with a specific expression, add .* to the expression, e.g. ".*\.mp3" will match all files that end with .mp3. While that may seem a bit redundant, it’s on par with “explicit is better than inplicit” (see The Zen of Python). See also: -I.


Name of the target file with references to regular expression groups caught in the source matches. References to groups are formed by a backslash character followed by he group number. Groups are indexed from 1. The group number can be contained within parentheses to disambiguate a reference followed by digits. Examples:


Automatic numbering can be introduced using a special \(index) reference. For example:

rename "IMG.*\.JPG" "Judy's Birthday \(index).jpg"

By default indexing starts with 1, increments with 1 and pads numbers with enough leading zeroes so that each filename uses the same amount of digits. This can be changed with the --index-first, --index-step, --index-digits and --index-pad-with options.

-I, or –case-insensitive

When used, regexes work in a case-insensitive manner, e.g. "lib" will behave like "[Ll][Ii][Bb]". Group references still hold the original case.

-l, or –lower

When used, renamed filenames are transformed to lower-case. This does not affect the source regex used (i.e. it still matches in a case-sensitive manner, unless -I is used). See also: -U.

-q, or –quiet

When used, no error output is given. The status of invocation should be determined via the return code.

-t, or –test

When used, the script will only fake renaming and verbosely state what it would do. Use this if you’re unsure of the effects your expression may cause.

-U, or –upper

When used, renamed filenames are transformed to upper-case. This does not affect the source regex used (i.e. it still matches in a case-sensitive manner, unless -I is used). See also: -l.

-v “except_regex”, or –except “except_regex”

When used, any filename matched by the original source regex will be also matched against the except_regex. In case there is a match, the filename is skipped. In other words, filenames that match except_regex will not be renamed.

The regular expression is local (e.g. writing "[0-3]" means “number 0-3 anywhere in the filename). This is to make the tool err on the side of caution by protecting from renaming too many files by accident when the user forgets to add dot-asterisk to -v. If you want to only match whole filenames, use the canonical global form (e.g. "^filename$"). See also: -I.


When using the special \(index) reference, this option specifies what number will the first index be. Default: --index-first=1.


When using the special \(index) reference, this option specifies what number will be added with each step to the first value. The specified number can be negative. Default: --index-step=1.


When using the special \(index) reference, this option specifies how many digits will be used in each reference. If a number has fewer digits, they will be prefixed by leading zeroes (or another character, see: --index-pad-with). A special value of auto can be used to automatically pad enough digits so that each filename has the same amount of them used. This is useful for ensuring your files will be sorted correctly even by dumb algorithms. Default: --index-digits=auto.


When using the special \(index) reference, this option specifies what character will be used for padding. Default: --index-pad-with=0.


Runs internal unit tests of all functionality. Does actual renaming in the temporary directory. Each test generates its own set of files.


This script requires Python 2.4+ with the argparse library. It can be used standalone or installed using pip or easy_install:

pip install rename
easy_install rename

Don’t have either of these? You can always grab the latest source release from the PyPI website or better yet equip yourself with easy_install by downloading and running


  1. The script will not let multiple files be renamed to a single name.

  2. The script will not let existing files to be overwritten.

  3. Both checks above are made for all matches before any renaming is performed.

  4. The script correctly preserves extended attributes and ACLs.

Other remarks

  1. Regular expressions supported by the script must conform to the syntax handled by Python’s re module.

  2. Actual renaming of a single file is done by the os.rename() function from Python’s standard library. No additional atomicity is ensured, e.g. if a single rename fails halfway through, the filesystem is left in a state of partially complete renaming.

  3. Due to differences in behaviour of different shells, the recommended form of execution is to put both arguments in quotation marks.

Possible future enhancements

  1. -s option to enable a “translate” mode to replace certain substrings with others. Proposed syntax:

    rename -s "substring_from" "substring_to" "file_match_regex"

    Example (translating underscores to spaces):

    rename -s "_" " " ".*\.txt"

    This would be more-less compatible with behaviour of the existing rename tool from the util-linux-ng package. One obvious difference would be that the file mask doesn’t use wildcards but regular expressions.

  2. -p option to create intermediate directories for the target. One tiny problem is maintaining atomicity of the whole transaction.

  3. -r option to make the source match recursive. Tricky to get right I guess, e.g. where to rename? Existing directory structure or new one?. Let the user decide? What’s the default? Etc. etc.

  4. Interactive mode. Things to be thought over: should the question appear before the transaction begins, before each step, or both? Should that be one option?


This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, version 3 of the License.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

DON’T PANIC. This code has been successfully used by its author and contains tests. However, be especially wary under these conditions:

  1. Renaming between filesystems.

  2. Renaming under non case-preserving filesystems.

  3. Renaming within very long paths.

  4. Renaming volatile state (e.g. rotating logs).

And if you do lose any data, it’s your fault. Have a nice day!


Script glued together by Łukasz Langa.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

rename-1.1.tar.gz (21.3 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page