LinuxQuestions.org - Find and Grep in Script

- Programming (https://www.linuxquestions.org/questions/programming-9/)

- - Find and Grep in Script - almost there (https://www.linuxquestions.org/questions/programming-9/find-and-grep-in-script-almost-there-550128/)

Find and Grep in Script - almost there

Hi - I need a bit of help with the final piece of a little bash script.

I need to traverse through a directory and its subdirectories looking for files that end with .txt. I need to grep through the found .txt files and print out a particular line, followed by the path and filename of the file itself.

For starters, from a terminal, when I do this:

Code:

find . -name *.txt | xargs grep metadate*

it almost works. I get the following output, as an example:

./directory/filename.txt:metadate20061212

What I need is for the metadate to come first, then the file location and name like this:

metadate20061212:./directory/filename.txt

Here is script I have so far:

Code:

#!/bin/sh

for i in $(find . -name *.txt); do

cat $1 | grep meta*;

echo $i;

done

This works, but it inserts a line break after the cat, so the output is like this:

metadate20061212
./directory/filename.txt

If someone could help me tweak the script so it outputs everything on one line I would be very appreciative. Or if there is a better way to do this, I would love to know it -- learning BASH never ends.

Thanks!

Assuming GNU grep on Linux:

Code:

sed 's/$.*$:$.*$/\2:\1/' <(grep -rF meta --include="*.txt" *)

or:

Code:

while IFS=: read a b;do echo "$b":"$a"; done< <(grep -rF meta --include="*.txt" *)

find . -name *.txt | xargs grep metadate* | awk -F ':' '{print $2 "\n" $1 "\n"}'

how about

Code:

find . -name "*.txt" -exec grep "metadate*" {} \;

--- rod.

Quote:

Originally Posted by Jackson1995

Code:

find . -name *.txt | xargs grep metadate*

Code:

#!/bin/sh

for i in $(find . -name *.txt); do

That will fail if there are any .txt files in the current directory; quote the pattern.

It will also fail if any filenames contain spaces.

Quote:

Code:

cat $1 | grep meta*;

That will also fail if any filenames contain spaces, or if any files in the current directory begin with "meta".

Quote:

Code:

echo $i;

done

Assuming GNU (and some other versions of) find and xargs:

Code:

find . -name '*.txt' -print0 | xargs grep "meta*" /dev/null | awk -F: '{ print $2 ":" $1 }'

Note that that will fail if any filenames contain a colon, or if a found line contains a colon.

something like this ?

Code:

for file in `find . -name '*.txt'`

do

    rslt=`grep 'meta*' $file 2>/dev/null`

    if [[ $? -eq  0 ]]

    then

        echo $rslt|awk -F: '{ print $2 ":" $1 }

    fi

done

something like this ?

Code:

for file in `find . -name '*.txt'`

do

    rslt=`grep 'meta*' $file 2>/dev/null`

    if [[ $? -eq  0 ]]

    then

        echo $rslt|awk -F: '{ print $2 ":" $1 }

    fi

done

EDIT: aaarrggghh; problem with my connection. Can some kind moderator remove this duplicate post.? Thx

for safety you should always quote when using the -name option

-name '*.txt'

so the glob isn't expanded by the shell for instance if you have a *.txt in your
current working directorry.

Wow, thanks for all the great responses. Clearly, I have a lot of Bash learning to do. :-)

Anyway, I ended up using radoulov's second piece of code:

Code:

while IFS=: read a b;do echo "$b":"$a"; done< <(grep -rF meta --include="*.txt" *)

and just piped that to a text file. Worked great.

Thanks again, everyone. This has been very helpful and educational.