1) find a string 'decor' inside files in a directory
2) get # of occurrence for decor for each file that has it

The first part of the problem is partially solved with find . -type f | grep -i decor *
But this doesn't seem grep for all files that are found from find but rather what's on top of current directory.
Do I have to use something other than * for grep to grep for files found from find?

what needs to be piped to so that it will display with a file name and the number of occurrences for decor for each file?


You can do:

find . -type f -exec grep -icH decor {} \; | grep -v :0\$

This will get the filename and count of each match and remove files with a count of 0 using the piped egrep

| improve this answer | |
  • No problem! Note the edit however, without the colon it'll ignore any number ending in 0, so 10, 100 etc... – arco444 Jan 14 '14 at 14:28
  • Just for the fun of it, and way beyond the scope of this exercise, I issued a find2perl . -type f -exec grep -icH {} \; to see if I'm willing to add the grep -v :0\$ part. Looks doable, perhaps later during Jeopardy. – Billy McCloskey Feb 2 '14 at 0:56

"Files in a directory" does not usually mean "and iside subdirectories recursively". I would just use

grep -c decor *

If you need to search subdirectories, too, you can use find, but connect the commands differently: pipe just connects output of the first command to the input of the second. Grep does not expect list of files on its input, though.

find . -type f -exec grep -cH decor {} + | grep -v :0
| improve this answer | |

You can use -exec option. find . -type f -exec grep decor {} \;

or use xargs find . -type f | xargs -i grep decor {}

| improve this answer | |
  • -1: i) this breaks on filenames with white space and ii) neither of the solutions return the number of matches for decor which is what the OP asked for. – terdon Jan 14 '14 at 15:31

TMTOWTDI - I like seeing the other variants, so here's mine as well:

$ find . -type f -exec egrep -li decor {} \;

$ perl -e '$c=0; while(<>){$c++ if /decor/;} continue {if(eof){print("$ARGV:$c\n"); close ARGV; $c=0;} }' *

Just because I felt like re-writing a Perl 1-liner that performs the same as an age old awk example taken from the pages of Kernighan and Pike.

| improve this answer | |
  • TMTOWTDI - Sounds like something worth doing. Note the perl2find not so useful suggestion, above - just for fun. You already worked out my answer, with the same reference. Bravo! – Billy McCloskey Feb 2 '14 at 1:02
  • BTW - That while(<>){} continue {if(eof){close ARGV;} idiom is worth memorizing. Again, that Perl idiom is worthy of memorization. – Billy McCloskey Feb 2 '14 at 1:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.