Exclude all subdirectories except one specific path in a find command

Question

I use a find command to search for all files with a given extension within the current subdirectories ignoring certain directory path:

find -L . \( -wholename "*/ignoredPath" -o -wholename "*/ignoredPath2" \) -prune -o -name "*.ext"

This is working pretty well but now I want to ignore another path except one precise subdirectory in it. How to do this without listing all the subdirectories to ignore?

For instance if I have the following folder tree:

*/base/pathToIgnore1
*/base/subdir/requiredPath
*/base/subdir/pathToIgnore2
*/base/subdir/pathToIgnore3
*/base/pathToIgnore4

How can I write the find command ignore all base subdirectories except the path */base/subdir/requiredPath?

I tried something like

find -L . \( -wholename "*/ignoredPath" -o -wholename "*/ignoredPath2" -o \( -wholename "*/base" -a ! -wholename "*/base/subdir/requiredPath" \) \) -prune -o -name "*.ext"

But it did not work, all the base subdirectories are ignored.

Answer 1

find -L . \
  \( -type d \
    \( -path "*/ignore" -o -path "*/indeed" -o \
      \( -path "*/subdir/*" ! -path "*/subdir/save" \
      \) \
    \) \
  \) \
  -prune -o -print

If you want extra filter (like -name "*.ext") you have to put that right before -print. The last part then looks like this

  -prune -o \( -name "*.ext" \) -print

Note that I changed the names for easier writeability and readability. Names starting with i are to be ignored. Names starting with s are to be shown. Names ending in file are files.

My tree looks like this:

$ find -printf "%y %p\n"
d .
d ./base
d ./base/subdir
d ./base/subdir/inform
f ./base/subdir/inform/imagefile
d ./base/subdir/isolate
f ./base/subdir/isolate/individualfile
f ./base/subdir/whatwhatinthefile
d ./base/subdir/save
f ./base/subdir/save/soundfile
f ./base/superfile
d ./base/indeed
f ./base/indeed/itemfile
d ./base/show
f ./base/show/startfile
d ./base/ignore
f ./base/ignore/importantfile

The output of the above command:

.
./base
./base/subdir
./base/subdir/whatwhatinthefile
./base/subdir/save
./base/subdir/save/soundfile
./base/superfile
./base/show
./base/show/startfile

Note that whatwhatinthefile in base/subdir. If you do not want the files in base/subdir then you have to exlude them explicitly. I tried but the command line became too ugly.

Depending on your use case it might be easier to define a shell function like this:

contrivedfind() {
  find -L . \
    \( -type d \
      \( -path "*/ignore" -o -path "*/indeed" -o -path "*/subdir" \
      \) \
    \) \
    -prune -o -print
  find -L ./base/subdir/save
}

Output is now:

.
./base
./base/superfile
./base/show
./base/show/startfile
./base/subdir/save
./base/subdir/save/soundfile

The only other thing different from before now is the missing ./base/subdir entry. But I think that does not matter in your case because you want to filter for files anyway.

As before you have to put any extra filter before the -print in the first find and this time also at the end of the second find.

Answer 2

You have to prevent that */base itself is matched:

Edit 1:

find -L . -type d \
  \( -wholename "*/ignoredPath" -o -wholename "*/ignoredPath2" -o \
  \( -wholename "*/base/*" -a \
  \( ! -wholename "*/base/subdir" -a ! -wholename "*/base/subdir/*" \) \) -o \
  \( -wholename "*/base/subdir/*" -a \
  \( ! -wholename "*/base/subdir/requiredPath" -a ! -wholename "*/base/subdir/requiredPath/*" \) \) \) \
  -prune -o -print

This is a bit finicking: It prevents accidental matches of directories like subdir2 which probably don't exist. But you never know... :-)