Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Suppose I have a directory with files and subdirectories underneath it. Some of the files contain a pattern "^File:" in their contents.

Is there some way I can use grep and a bash command to remove files that contain this pattern? Note I am talking about contents of the file, not the filenames.

share|improve this question
This question might be better for – Hemm Oct 27 '11 at 23:37
If you could narrow your search by a filename pattern, then you'd get a lot better efficiency from any scripts you get. – Chris Ting Oct 27 '11 at 23:43
up vote 2 down vote accepted

A simple piped find should suffice (caveat emptor, I haven't tested this against a large set of data -- backup anything important):

find /path/to/dir -type f -exec egrep -Il '^File:' {} \;|xargs rm -fv
share|improve this answer
@Hemm -- Actually I tested mine against files with spaces and it removes them just fine. You probably want to add -I to your grep or he'll be deleting binary files too. – Garrett Oct 28 '11 at 0:24
@Hemm's example can also potentially run into argument limitations of grep if too many files are matched by the wildcard. – Garrett Oct 28 '11 at 0:33
I tested mine on 100k+ files containing the string with no problems. And I already have -l in it. – Hemm Oct 28 '11 at 1:54
I as in India. man grep ;) – Garrett Oct 28 '11 at 2:31
-I- blame sans serif. – Hemm Oct 28 '11 at 2:59
grep -rlI "meow" * | xargs -I{} rm -v {} 

The above code removes files containing meow.

This also handles cases where there are filenames with spaces and other characters. xargs rm -f alone, as suggested in other answers, will fail.

share|improve this answer
That doesn't descend into subdirectories. – Keith Thompson Oct 28 '11 at 0:07
Whoops. -r added. – Hemm Oct 28 '11 at 0:14

Here's a one-liner.

grep -r "^File:" /path/to/directory | cut -d: -f1 | sort | uniq | xargs rm -f
share|improve this answer
Much more verbose than it needs to be. Use grep's -l option to list just the file names. The sort and uniq steps are unnecessary, since the files are just going to be deleted. grep -r -l "^File:" /path/to/directory | xargs rm -f. Or, if some of the file names might contain funny characters (particularly spaces): grep -r -l "^File:" /path/to/directory | tr '\n' '\0' | xargs -0 rm -f. (I have not tested these; try xargs echo rm -f first to make sure it's generating the right list of files.) – Keith Thompson Oct 28 '11 at 0:06
@ChrisTing: Could you provide some explanation about what your code does? – Tom Wijsman Nov 1 '11 at 1:34
@KeitThompson: Could you please consider to copy that in an own answer? :) – Tom Wijsman Nov 1 '11 at 1:35

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .