Shell Programming and Scripting

Hi,

I have used this

Code:

 find . -type f -exec grep "491629461544"  {} \; > test.txt

to find the string and export it to file.

The problem is, when i used the above command in Shell script, the file is created but the searched string is recursively written in that file and even the file size exceeds 1 GB.

so, how to write the obtained string only once in newly created file.


Please Suggest me.

assuming I understand:

Code:

find . -type f -exec grep "491629461544"  {} \; | head -1 | > test.txt

I'm only curious, is there a difference between

Code:

find . -type f -exec grep "491629461544"  {} \; | head -1 | > test.txt

and

Code:

find . -type f -exec grep "491629461544"  {} \; | head -1 > test.txt

(without the last pipe)

Does this even work?
I would think that the output of head is piped to a null process which doesn't output anything and thus test.txt should be empty.


--- edited out --- My head was in the sand the other night and the question I posted was just wrong.

Still I'm curious WHY that find didn't stop and test.txt is filled beyond belief. find is working on the current dir, in which test.txt is created as well. So grep could read that file, match the string in it, and output to it ad nauseam. But - at least my - grep refuses to do so:

Code:

grep: input file `./test.txt' is also the output

even if executed in a script. I'd like to see more of that script.

It seems that grep is a built-in to your shell. There's no way otherwise that it would know that the file to search was also the target of redirection.

I just executed the original example find across a small set of files from which only one line out of the files matched the pattern. The output ended up with two lines. Adding the -H parameter to grep shows that the first line in the output file comes from the original file, and the second is tagged with the name of the output file.

---------- Post updated at 10:55 ---------- Previous update was at 10:51 ----------

I think this might be a better solution -- more efficient in terms of processes.

Code:

find . -type f  ! -name test.txt |  xargs grep "491629461544"  >test.txt

I was thinking about this as well, but -

Code:

grep --v
grep (GNU grep) 2.10

- it's a file in /bin and not a built-in. It must be using functionality like lsof to determineits own open files in the target directory?!
I could reproduce what you posted by adding a pipe to cat; now the output file contains what you post, i.e. one line referring to itself.

I still wonder why nanthagopal's disk fills up, though ...