grep: --exclude-dir that contains some text in the name

Question

I need to use:

grep -rio pattern --exclude-dir=".config/dir1*" --exclude-dir=".test*"

in order to grep for that pattern avoiding all the folders that contain .config/dir1 and .test in the name,

E.g. .config/dir1, .config/dir1/dir2, .test and ./other/.test should be excluded.

What I did does not exclude those folders, could you please help me?

Thanks

Answer 1

Your command does exclude files in .test/ and in other/.test/.

It doesn't exclude files in .config/dir1 because --exclude-dir only looks at the base name of the directory, not at its full path. You would need --exclude-dir="dir1", which excludes elsewhere/dir1 as well. Note that if a directory is excluded, its subdirectories are excluded as well, so --exclude-dir="dir1" also excludes files in .config/dir1/dir2.

If grep's exclude patterns aren't sufficient for you, you can combine find with grep. The syntax of find is more complex, but it's more powerful. For example, you can exclude .config/dir1 while not excluding other directories called dir1. (I assume that you have GNU find since you have GNU grep.)

find . \
     -path "./.config/dir1" -prune -o \
     -name ".test" -prune -o \
     -type f -print0 |
  xargs -0 grep -io pattern