I have a directory hierarchy that contains numerous .gz files. I want to be able to recursively grep them for the string "foo". From what I've read online the following should work:
zgrep -R -H "foo" .
However, this never returns any results. If I replace the dot with the name of a file it does work. For example,
zgrep -R -H "foo" myFile.gz
however, obviously, this no longer will be recursive.
I know "foo" is in some of the files because the following command returns many results:
find . -iname "*.gz" | xargs zgrep "output" | less
Does anyone know why my recursive zgrep command is not working. I'm on a RHEL linux box
5 Answers
The way I usually do is:
zgrep "foo" $(find . -name "*.gz")
or (however, the file name will be printed before each result instead of each line --not just the files with matches--):
find . -name "*.gz" -print -exec zgrep "foo" {} \;
If that command returns "Argument list too long", try this way:
for I in $(find . -name "*.gz"); do zgrep "foo" $I; done
You have to install zutils. This will replace the default and limited zgrep on your system with a recursive-able one.
On debian based systems you run apt-get install zutils then you can zgrep -rH myword . and use most of the other parameters from grep you know and love.
Your almost there. Try this:
zgrep -R -H "foo" *.gz
EDIT: Hmmmm.... intriguing!
According to my zgrep, -R (Recursive) is not an option. Its simply not supported. Id have a check to see what the man page of your zgrep says.
One alternative, which depends on only one level of subdirectories is to do this:
zcat */*.gz | grep <needle>
But I would suggest that your find command is probably better!
Use sift - it's super-fast.
sift -z foo
-z is for unpacking gzips, it descends into subdirectories by default.
Shows filenames also by default (your -H option).
PS. Thanks to your question I discovered this tool today. Just battletested it, apart from the functionality you need, I have a 5x speedup on some production logs when comparing to zgrep.
it also works with:
zgrep "foo" */ *.gz
