Suppose I have a tarball (.tar.gz file) which is large and I only want to extract a specific file from it. If I know the name of the file all I have to do is pass the file’s relative path that it is stored under to the command line.

Here is an example of the error you will get if you pass the incorrect file specification:

$ tar zxvf dirtree-tarball.tar.gz file-7-30003.txt
tar: file-7-30003.txt: Not found in archive

Since I don’t have the full path, I can just search for it:

You can use tar to copy a directory tree to another directory in one shot along with preserving ownership, permissions and timestamps. This also avoids making an intermediary tarfile which may cause problems if the size of the file copy is large and the storage resources are low. Just cd to the top of the directory that you want to copy and begin.

Let’s assume that you want to copy the contents of the source directory to a target directory:

If you have a lot of files in a directory and you only need to tar a subset of them you can create a list of the files you want in a text file and pass it to the tar command like this:

$ tar -cvf tarball.tar -T filelist.txt

or

$ tar cvf tarball.tar $(cat filelist.txt)

If you want to create a tar archive of only the files of a directory and exclude any subdirectories you can use the ls -la command and pipe the output to awk. However you need to remove the first 8 fields from the output and leave all of the remaining parts of the line in case there are spaces in the filename. One quick and dirty way of doing that is to set each of the 8 fields to a blank and then use sed to trim the leading spaces. You can optionally add quotation marks around the filename in your output too.

Sometimes you may need to copy an entire directory structure to another system using the command line. Here is a quick way to do it using the tar command:

cat myfile.tgz | ssh user@host “tar xzf - -C /some/dir”