19

Depending on how a zip file is created, sometimes it will extract all of the files directly, and sometimes it will extract the files into a subdirectory.

If the latter is true, how can I force the unzip command to "ignore" that first level directory?

Example:

cd /tmp
wget http://omeka.org/files/omeka-1.5.1.zip
mkdir omeka
unzip omeka-1.5.1.zip -d omeka/
cd omeka/
ll

What I'm getting is /tmp/omeka/omeka-1.5.1/:

total 12
drwxr-xr-x 3 root root 4096 2012-05-08 18:44 ./
drwxrwxrwt 6 root root 4096 2012-05-08 18:44 ../
drwxr-xr-x 5 root root 4096 2012-04-20 14:54 omeka-1.5.1/

What I want is of the files extracted to /tmp/omeka/, (one level up and no version number included in the directory structure)

/tmp/omeka/(files)

I know I can use the -j option to "junk paths" but I want to keep the subdirectory structure, just not the top level directory structure. How can I do this?

  • ps: I'm aware that I can extract and then mv the files, but I wanted to see if there was a way to do this straight away from the unzip command – cwd May 8 '12 at 18:55
  • I'm not aware of one. But if the filename always matches (apart from the extension) the directory name, it's pretty straightforward to script the rename of the extracted dir. – Mat May 8 '12 at 19:07
  • 3
    I see what you're trying to do, but I wouldn't recommend it. Stripping the version number from your project sources is a bad idea; you should just extract normally and use a ln -s if you want a shorter name without version information. That way you always know what version you are running and can easily update to a new version by switching the link. – Burton Samograd May 8 '12 at 19:38
  • @BurtonSamograd - that would be ideal, but I'm still curious if there is an effective way to do this because some software, take Wordpress as an example, puts every version inside a /wordpress/ directory (no version number) inside the zip file. It is indeed fine to unzip and then mv but having no control of this and having to do it in two steps has always gotten on my nerves a little. Fortunately Wordpress also comes in a .tar.gz flavor :) – cwd May 9 '12 at 2:01
  • Try cd /tmp/omeka && ln -s -T . omeka-1.5.1 – James Youngman May 9 '12 at 22:39
4

Use a FUSE filesystem that allows you to browse archives like directories, such as AVFS. Use cp to extract the files to the directory of your choice.

mountavfs
cp -Rp ~/.avfs/tmp/omeka-1.5.1.zip\#/omeka-1.5.1 omeka

Since we're assuming that there is a single toplevel directory in the archive, you can shorten this to

cp -Rp ~/.avfs/tmp/omeka-1.5.1.zip\#/* omeka
  • @Giles you are the expert who comes and answers all the questions nobody else can answer, so I guess this is truly the only way to do this. I was hoping unzip had some magic to do this, such as specifying the files to be extracted like */*, that I could use without installing another utility, but I guess not. Thanks. +1 – cwd May 9 '12 at 2:05
6

If your zip file contains no directory structure or you do not need to preserve it, you can use this:

cd /tmp
wget http://omeka.org/files/omeka-1.5.1.zip
unzip -j omeka-1.5.1.zip -d omeka
cd omeka
ll
3

This script is not robust, but works in the simple cases:

...
dest=omeka
unzip omeka-1.5.1.zip -d $dest/

if [ `ls $dest | wc -l` == 1 ]; then
  subdir=`ls $dest`
  mv $dest/$subdir/* $dest/
  rmdir $dest/$subdir
fi

It just checks to see if there is exactly one subdirectory, and if so, moves everything up out of it then deletes it.

2

I've just registered today, so I can't upvote @SteveBennet answer and can't add a comment there.

Based on his answer I created a recursive function like this:

...

shopt -s dotglob # To include hidden files in the move command

function moveSub {
  local dest=$1
  if [ `ls $dest | wc -l` == 1 ]; then
    local subdir=`ls $dest`
    moveSub "$dest/$subdir"
    mv $dest/$subdir/* $dest/
    rmdir $dest/$subdir
  fi
}

moveSub "$dest"

Just like @SteveBennet said: this script is not robust, but works in the simple cases.

Hope it's useful.

0

Need looks like patch --strip=n option

It would be desirable to have an option like patch --strip=number (or -pnumber) that chops number path components at beginning (ref. Comparing and Merging Files: patch Directories).

Other solution, applicability

Failing that, there is another solution, kind of hackish but works. In my case I wanted to overwrite a directory full of existing files with the content of a zip, without extracting the full zip first.

When this solution works and @Gilles's solution is available, the latter will work, too. Yet this solution does not require availability of a FUSE filesystem.

How to do it

Assuming you know the extra directory, in you case omeka-1.5.1, you can do this:

mkdir omeka
ln -s . omeka/omeka-1.5.1 # create a symlink that redirects output
unzip omeka-1.5.1.zip -d omeka/
rm omeka/omeka-1.5.1 # remote symlink

unzip will try to unzip to omeka-1.5.1 which is actually a symlink to containing dir. As a result, files will land in omeka directly.

Possible variants

You can imagine variants to redirect one or more parts of a depeer hierarchy.

ln -s ../myfoo omeka/omeka-1.5.1/foo
ln -s ../../mybarxyzzy omeka/omeka-1.5.1/subdir/bar

Conclusion

This solution is somewhat specific, yet it has its use cases and I just happened to use it.

0

Most likely you want to run unzip with the -j option, as in:

unzip -j files.zip -d ./output/

But you might be surprised by its behavior:

-j junk paths. The archive's directory structure is not recreated; all files are deposited in the extraction directory (by default, the current one).

If you extract files.zip in output directory ./output and you get this as a result:

./output/files/subfolder1/f1
./output/files/subfolder1/f2

then the -j option will give you this (all paths striped):

./output/f1
./output/f2

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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