Further processing of find -exec

grob115 · 02-12-2011, 05:46 AM

If I want to do a grep on the files on the whole server I can do the following:

Code:

find / -type f -exec grep '<what I want to grep>' '{}' \;

What if I want to pipe the "grepped" output for further processing? I can't just simply do this. It just doesn't work.

Code:

find / -type f -exec grep '<what I want to grep>' '{}' \; | grep '2nd level of grep'

How about I want to find the number of occurrences per file returned? For example, if the output of the first command above returns...

Code:

file 1: ad;flkjasd;lfk <what I want to grep> asdfl;kj
file 1: ad;flkjasd;lfk <what I want to grep> asdfl;kj
file 2: ad;flkjasd;lfk <what I want to grep> asdfl;kj
file 3: ad;flkjasd;lfk <what I want to grep> asdfl;kj
file 4: ad;flkjasd;lfk <what I want to grep> asdfl;kj

How do I structure the command so I can see something like?

Code:

file 1: 2
file 2: 1
file 3: 1
file 4: 1

crts · 02-12-2011, 09:11 AM

Hi,

for the problem that you described try this first:

Code:

find . -type f -exec grep -cH 'pattern' '{}' \;

Then check this thread for other possibilities to further process the results:
http://www.linuxquestions.org/questi...o-slow-816525/

Finally, read the man-page for more info on how you can modify the output of grep. Chances are that you do not need the technique described in the linked post.

Hope this helps.

someshpr · 02-12-2011, 09:27 AM

grep has a recursive option. So probably you can get rid of the find command if you wish. Something like this should give you the count of occurrences of your search pattern on all files in current directory and all its subdirectories:

Code:

grep -rc '<what I want to grep>' .