How can we GZip every file separately?
I don't want to have all of the files in a big tar.
Add a comment
|
8 Answers
You can use gzip *
Note:
- This will zip each file individually and DELETE the original.
- Use
-k(--keep) option to keep the original files. - This may not work if you have a huge number of files due to limits of the shell
- To run gzip in parallel see @MarkSetchell's answer below.
1 Comment
MCunha98
If I want to run it under a crontab, which command line I'll put ? For example /tmp/app/gzip * ?
Easy and very fast answer that will use all your CPU cores in parallel:
parallel gzip ::: *
GNU Parallel is a fantastic tool that should be used far more in this world where CPUs are only getting more cores rather than more speed. There are loads of examples that we would all do well to take 10 minutes to read... here
3 Comments
zx8754
Knew about parallel, but keep forgetting to use it! Ran the accepted answer then scrolled down to see your answer... saved lots of hours! Maybe a good idea to make your comment an edit to your answer?
Californian
Is there a way to use parallel if the argument list is too long? Have about 60k files that I need individually compressed.
Mark Setchell
@Californian Sure
find . -name XYZ -print0 | parallel -0 gzipAfter seven years, this highly upvoted comment still doesn't have its own full-fledged answer, so I'm promoting it now:
gzip -r .
This has two advantages over the currently accepted answer: it works recursively if there are any subdirectories, and it won't fail from Argument list too long if the number of files is very large.
2 Comments
Arlen Beiler
Does this keep or delete the files? Do you still need to add the -k option mentioned in the other answer.
leekaiinthesky
Add
-k if you want to keep the original files.If you want to gzip every file recursively, you could use find piped to xargs:
$ find . -type f -print0 | xargs -0r gzip
4 Comments
Idelic
No need for find+xargs. Gzip can handle recursion itself:
gzip -9r .
musiphil
As always,
find . -type f -print0 | xargs -0r gzip is better.musiphil
For the equivalent of
gzip *, you may also need -maxdepth 1 in find.
Olivier Cailloux
@musiphil: good point about protecting for spaces in file names! I just edited the answer to integrate your comment (waiting for peer review).
Try a loop
$ for file in *; do gzip "$file"; done
Comments
Or, if you have pigz (gzip utility that parallelizes compression over multiple processors and cores)
pigz *
Comments
The following command can run multiple times inside a directory (without "already has .gz suffix" warnings) to gzip whatever is not already gzipped.
find . -maxdepth 1 -type f ! -name '*.gz' -exec gzip "{}" \;
A more useful example of utilizing find is when you want to gzip rolling logs. E.g. you want every day or every month to gzip rolled logs but not current logs.
# Considering that current logs end in .log and
# rolled logs end in .log.[yyyy-mm-dd] or .log.[number]
find . -maxdepth 1 -type f ! -name '*.gz' ! -name '*.log' -exec gzip "{}" \;
Comments
fyi, this will help to overwrite if an existing gz file along with creating few gz file if it is not present:
find . -type f | grep "in case any specific" | grep -E -v "*.gz$" | xargs -n1 -P8 sh -c 'yes | gzip --force --best -f $0'
