Applying a script to all the files of the find command

flamingo_l · 10-18-2010, 07:47 AM

Code:

find "$SOURCEDIR" -type f -name "*$ITEM" -printf "%P\n"

I want to apply some shell script to the files outputed by the find command.

How can i do this.? Can anybody help me on this.

There are multiple files directories and also multiple files.

genderbender · 10-18-2010, 07:54 AM

Isn't exec able to do that? I always use the following to change hostnames on unix systems

Code:

vi `find . -type f -exec grep -li THEOLDHOSTNAME {} +`

The exec could call a bash script I suspect.

GrapefruiTgirl · 10-18-2010, 07:54 AM

Code:

find "$SOURCEDIR" -type f -name "*$ITEM" -printf "%P\n" | xargs -I {} /path/to/your/script {}

`find` is recursive, so unless you tell it otherwise, it will find all files and all folders under the search path. This command I give, will execute:

Code:

/your/script <find result>

for every result from `find`.

If this doesn't help or doesn't work as you expect, tell us what it does or doesn't do correctly.

GrapefruiTgirl · 10-18-2010, 07:57 AM

Plus - genderbender is probably right too, as far as I know; you could use `find`'s -exec command too. I think it would work.

H_TeXMeX_H · 10-18-2010, 08:19 AM

Yes, but remember that using -exec will run that command on each file individually, while xargs will run once (or a few times) on all files.

EDIT:
One more thing, I would use the find -print0 and xargs -0 options to help with possible white space issues.

grail · 10-18-2010, 08:22 AM

Which then boils down to how does your script handle input, ie can it take more than a single argument?

David the H. · 10-18-2010, 08:36 AM

When you use a plus sign at the end of the -exec expression, it acts similarly to xargs, concatenating all the file names into as few runs as possible; as opposed to using a semicolon, which runs the command separately for each file found. See the find man page for details.

Your script/program must be able to handle the multiple inputs, of course.

flamingo_l · 10-19-2010, 12:24 AM

Thanks for all your replies.

The following is the script which needs to be run on the output files of the above find result.

Quote:

awk -f Extract.sh <filename> |awk '{sub(/^[ /*]+/, ""); print}'|awk -f Format.sh

Where Extract.sh and Format.sh are shell scripts.

Meanwhile, i will try the different options provided by all of you.

grail · 10-19-2010, 12:57 AM

Well without seeing the other scripts I would have thought it might be time to merge the 3 scripts to get done what you require.

Seeing where you need the filename I am not sure if any of the above would be correct (I could be well wrong here)
I would probably suggest a while loop getting input from the find and performing your tasks as based on post #8 what you have will only act on separate files.

flamingo_l · 10-19-2010, 02:51 AM

Yes Grail.

The same thing i have done and got it.

Here is the code:

Quote:

while read file; do
echo "File name: " $file
awk -f Extract.sh $SOURCEDIR/$file|awk '{sub(/^[ /*]+/, ""); print}'|awk -f Format.sh
done < <(find "$SOURCEDIR" -type f -name "*$ITEM" -printf "%P\n")

Posted this thread to know some other way of doing this.

Also,

Is there any way to recognize if there is no output from the below:

Quote:

awk -f Extract.sh <filename>

grail · 10-19-2010, 06:13 AM

Quote:

Is there any way to recognize if there is no output from the below:

Quote:
awk -f Extract.sh <filename>

Would need more information on what Extract.sh does?

flamingo_l · 10-19-2010, 08:40 AM

hi Grail,

The following is the script.

Quote:

{
if ($0 ~ "START")
{
#print $0;
getline;
do
{
print $0;
getline;
}while ($0 !~ "END");
print $0
}
}

It will extract the text between the START and the END tags in a given input file.

grail · 10-19-2010, 09:25 AM

Well I am not sure the last print $0 should be there as it will print line containing END??
Maybe you could use this much simpler version:

Code:

/START|END/{f=!f;next}f

Maybe if you show me some input and what is in Format.sh we could make it a lot simpler?

flamingo_l · 10-20-2010, 12:39 AM

hi Grail,

Yes print $0 is required to print the line END.

How can i execute the code given by you.?

My Inout file may or may not contain the START or END tags, so i want to skip the file which is not having START and END tags. Is there any way to recognize this?

It is the same script given by you in the following thread.

http://www.linuxquestions.org/questi...t-file-837330/

grail · 10-20-2010, 07:03 AM

So why do we need the first 2 awk scripts?

If you wish to test the above just put it into your extract.awk file and run as you have above.