I have a 38GB folder with 800 MP4 videos in it. After re-downloading it, the file name has no spacing, and all words are joined, but it's still TitleCase.

So from TitleCase I need Title Case.

What would be the most effective way of bulk renaming these files?

I remember a rename or autorename being included a long time ago in my distro, but I don't seem to have it now.

  • 1
    Unpopular opinion: if renaming one file manually takes 10 seconds, that's about 130 minutes, then ten minute total per day will get the job done within two weeks. Assuming it's not urgent.
    – shuhalo
    Dec 5, 2022 at 10:54
  • 6
    Can you tell us exactly which change you want to perform? E.g. all "ThreeJoinedWords" ->"Three Joined Words"? (Tongue-in-cheek: Best would be to give us this information as an sed command!) By the way, if my understanding is remotely correct: I typically perform the opposite name transformation, replacing all funny chars, including whitespace, by underlines or such. I would strongly advise against file names with spaces. The reason becomes obvious if later the need arises to change your names with spaces again. Dec 5, 2022 at 11:15
  • 6
    Please edit your question and i) tell us your operating system, ii) if it is a Linux, tell us which one, iii) give us some examples of your input file names and iv) the output file names you would expect from that example. Finally: please, please, please don't make the mistake of adding spaces to file names if they don't already have them. That will only make every future manipulation of the files more complicated for little benefit. You can use _ instead of spaces and they will still be easily readable while not being harder to handle.
    – terdon
    Dec 5, 2022 at 12:20
  • 1
    @terdon - fwiw, if it's about renaming CamelCase to space-separated words then this is a dupe of Bulk renaming of camelcase files to include spaces Dec 5, 2022 at 20:45
  • 1
    A little suggestion: Since the files are already on a Unix and/or Linux box, why not leave the spaces out and forget about having to escape them when addressing/accessing/performing an operation on any of the files? Just a thought.
    – B.Kaatz
    Dec 6, 2022 at 3:02

4 Answers 4


If you want to add spaces between each 'Words' of mp4 filename TitleCase (PascalCase to Word Separated By Spaces):

rename -n 's/\B[[:upper:]]/ $&/g' ./*.mp4


rename(./FooBarBaz.mp4, ./Foo Bar Baz.mp4)

Check this post that details in depth the Perl's rename

What about rename different versions and usage ? What is the recommended way to use the Perl version especially?

  • I wouldn't call that particular rename implementation (there are several perl-based renames these days) reliable, especially when called like that. Try for instance after touch './--e=system"reboot"#.mp4' Dec 5, 2022 at 17:18
  • 1
    As mentioned earlier, the perl-rename on Arch is a different implementation with a different API. Dec 22, 2022 at 23:02
  • 2
    @GillesQuenot posting a link to your answer on every question or answer that mentions perl rename is, at the very least, verging on spam. IMO, it's way beyond "verging" and definitely crossed over to spam territory - you've posted the exact same comment to three of my (ancient) answers so far.
    – cas
    Dec 26, 2022 at 2:21
  • 1
    Please stop leaving comments linking to this just because an answer mentions rename. That really is kind of spammy. I don't know why you chose to add all this detail to an answer of a closed question which did not ask about perl rename in the first place, but if you want to link people to it, please only do so when it is actually relevant to what a question asked. Commenting on answers which just mention rename isn't really helpful.
    – terdon
    Dec 28, 2022 at 9:55
  • 2
    I'm not offended and I'm not taking it personally. I just hate spam - with a deep and intense and eternal loathing. Getting 3 (so far) notifications of the exact same self-promotional comment is spam, and that's just the notifications I got - I don't know who else you've spammed, or how many. If I were to get any more, it would be even worse because it would be more spam (and since I've posted numerous answers using perl rename over the last decade or so, if you continue your spammy ways, I'll almost certainly get more).
    – cas
    Dec 28, 2022 at 10:55

I don't know what would be the most efficient way (I think you meant efficient), but I would quickly write a for loop like:

for file in *.mp4; do
  newname="$(echo "$file" | sed 's/\(.\)\([A-Z]\)/\1 \2/g')"
  mv "${file}" "${newname}"


  newname="$(echo "$file" | sed 's/\(.\)\([A-Z]\)/\1 \2/g')"
#        ^-------------------- Assign to  variable "newname" value…
#         ^------------------- "$()": as output by commannd in parentheses;
#                              use "" to avoid word splitting


echo $file | sed 's/\(.\)\([A-Z]\)/\1 \2/g'
#    ^-------------------------------------- output old file name
#          ^-------------------------------- pipe to `sed` command

sed is the name of the "stream editor"; it takes input, executes a command on it, and produces output. Here, the command is s, as in "search and replace".

s/\(.\)\([A-Z]\)/\1 \2/g
^^ ^  ^ ^      ^ ^  ^  ^
|| |^ | | ^^^  | |  |  |
\------------------------ s: search and replace
 \----------------------- /: Set the search;replace;flags separator to "/"
   || | | \|/  | \  /  |
   \--+------------------ \(…\): a "capture group" (the first one);
    |   |  |   |  ||   |         whatever is matches the content will be
    |   |  |   |  \|   |         available as \1
    \-------------------- .: We match ".", which means *any* character
        |  |   |   |   |  (which precludes this from matching at start of line)
        \------+--------- \(…\): Second capture group, \2
           \------------- [A-Z]: Match any capital letter
                   |   |
                   \----- Replacement: "\1 \2" replace
                       |  "characterbeforecapitalletter""Capitalletter" with
                       |  "characterbeforecapitalletter" "Capitalletter"
                       \- g: Flag that means "global": Repeat this until
                             end of line
  • Is the choice of this particular regex robust against the presence of newline characters (\n) in the filename?
    – Hastur
    Dec 5, 2022 at 7:42
  • 1
    Yes. Sed would operate on both lines separately and not care (side from doing the right thing of not inserting a space at the beginning of the line) , and the file and newname variables works continue to contain the newline character, @Hastur Dec 5, 2022 at 8:25
  • 2
    Oh that diagram in the last code block!
    – jcaron
    Dec 5, 2022 at 16:54
  • 1
    @jcaron sorry about that, I could not resist making it! Dec 5, 2022 at 16:58

To separate out words in title-case, like for FooBarBaz.mp4 to become Foo Bar Baz.mp4, you can do insert a space before every uppercase letter except when it's the first character in the file name, which with zsh's zmv you could do (recursively) with:

$ autoload -Uz zmv
$ zmv -n '(**/)(?)(*.mp4)' '$1$2${3//(#m)[[:upper:]]/ $MATCH}'
mv -- FooBarBaz.mp4 'Foo Bar Baz.mp4'
mv -- OnceUponATime.mp4 'Once Upon A Time.mp4'

(remove the -n for dry-run if happy).

Beware it changes Foo-Bar.mp4 to Foo- Bar.mp4.

zmv -n '(**/)(*.mp4)' '$1${2//(#b)([[:alpha:]])([[:upper:]])/$match[1] $match[2]}'

Would only insert spaces between a letter and an uppercase letter, but would not work for OnceUponATime above as the space would be inserted between n and A, but not between A and T as the A would have already be consumed by the previous substitution.

As zsh globs don't have the equivalent of perl's look around operators, working around that is more difficult. A simple approach in this case though is to just repeat the substitution an extra time:

$ zmv -n '(**/)(*.mp4)' '$1${${2//(#b)([[:alpha:]])([[:upper:]])/$match[1] $match[2]}//(#b)([[:alpha:]])([[:upper:]])/$match[1] $match[2]}'
mv -- ABCDEF.mp4 'A B C D E F.mp4'
mv -- AChristmasCarol.mp4 'A Christmas Carol.mp4'
mv -- FooBarBaz.mp4 'Foo Bar Baz.mp4'
mv -- LeSongeD\'UneNuitD\'Été.mp4 'Le Songe D'\''Une Nuit D'\'Été.mp4
mv -- LifeOfΠ.mp4 'Life Of Π.mp4'
mv -- OnceUponATime.mp4 'Once Upon A Time.mp4'

Other answers have described tools that allow the change based on regexes and that is a fine approach. However, those answers do not address an important point (IMHO): preview capability before actually making changes on disk (that may be hard to reverse).

Enter Emacs and its dired mode, in particular wdired, documented in the manual. One would open dired for the directory containing the files, toggle wdired to make the file names editable, make changes - most likely using interactive regex search and replace, review the changes, and finally commit them to disk.

I am aware that installing Emacs for this one use case is a heavy-handed approach, but I would point out that it handles file names with difficult characters, e.g. spaces, trivially.

  • 1
    preview capability offered by me and Stéphane Chazelas Dec 5, 2022 at 18:38

Not the answer you're looking for? Browse other questions tagged or ask your own question.