I'm trying to archive files with tar and ignoring files larger than 100MB. Is that possible?
PS: Not have to use tar.
2 Answers
Yeah there are probably a few more ways of doing this but you're probably best off with find
as it's so tunable.
find /path/to/dir -size -100M | xargs tar cvf archive.tar
or
tar cvf archive.tar $(find /path/to/dir -size -100M)
If there are other directories in the current folder which you do not want to archive, then use:
find /path/to/dir -maxdepth 1 -size -100M | xargs tar cvf archive.tar
Per the comments below, spaces can throw out some errors but in my testing they still seemed to get added to the archive when using the first method (a little strange in itself) but if you're worried, you can fix the find command to replace spaces with escaped versions.
As ever, there are a billion ways of doing this.
find /path/ -size -100M | sed 's/ /\\ /g' | xargs tar cvf archive.tar
find /path/ -size -100M -exec echo '"{}"' \; | xargs tar cvf archive.tar
-
2Both options could give problems if the OP has filenames containing spaces.– enzotibFeb 21, 2012 at 12:20
-
... agreed - will need to use
-print0
in the find command as well as--null
for the tar command to cope with this scenario.– fossfreedom ♦Feb 21, 2012 at 13:12 -
While creating a huge archive
tar cvf myfile.tar
command fails. I strongly suggest use totar rvf myfile.tar
command. Feb 21, 2012 at 16:21 -
In case someone uses this to filter out files of smaller size, it might be useful to read this first: unix.stackexchange.com/questions/275925/… . TL;DR
-size -1k
doesn't work as expected, and might select only files that have a size of 0. You might need to use! -size +1k
. (Same for1M
,1G
, etc.) Aug 31, 2021 at 13:26 -
And in case you use
xargs
, make sure to add-d '\n'
orxargs
will break iffind
finds files with whitespaces in their names. See stackoverflow.com/questions/16758525/… Aug 31, 2021 at 13:51
Just exclude the big files like this:
tar cf archive.tar /path/to/archive --exclude-from <(find /path/to/archive -size +100M)