Shell Programming and Scripting

I need to construct a command that finds directories which contains more than n matches of a certain filename.

E.g. I have many directories at different locations and want to find all directories that has 2 or more .dat-files.

I thought of using find and maybe the exec parameter to issue an ls|wc -l, but that would just return the number of files, not the directory itself (that i need).

Any suggestions?

Hi, welcome to the forum. You could use an awk command to cut off after the last / and count the find output, for example:

Code:

find . -name '*.dat' | awk -F/ '{NF--}++A[$0]==n' OFS=/ n=2

Hi Scrutinizer,

I little new to UNIX.. Could you please explain the below awk command posted by you in detail ???
Thanks in advance!
Vasanth.

Sure:

(Remove the wrong statement.)

Below code will be easy for beginner for understanding.

Code:

find . -name '*.dat' -type f -exec dirname {} \; |sort |uniq -c |awk '$1>n' n=2

No, I think my code is correct. It needs to be == , otherwise the same directory gets printed multiple times. Put in another way, once the number of occurrences reaches 2, it answers the criterion of 2 or more .dat files and the directory should be printed, no matter how many more .dat files are found in the lines that follow..