[SOLVED] How to use find with two lines together?

guessity · 01-19-2010, 03:00 AM

I am trying to search for

Code:

<cats>
Sports
</cats>

how do i use the find command to search for <cats>Sports

The problem here is that the word Sports is in the next line of <cats> so I cant use -

Code:

find /test/upload/ -type f -name "*.xml" -exec grep -l "<cats>Sports" {} \; -print

How do I put the next line string? plz help..

bigrigdriver · 01-19-2010, 04:21 AM

Try this:

Code:

find /test/upload/ -type f -name "*.xml" -exec grep -l -A 1 -B 1 "Sports" {} \; -print

-A 1 print one line After the line containing Sports
-B 1 print one line Before the line containing Sports

guessity · 01-19-2010, 04:49 AM

thx but the file might contains a reference of word Sports anywhere so that command showed me all the files which ever contained the word sports irrespective of where the word is present.

the only unique reference is when sports is within <cats>

I just wanted to see the files which contains word withing <cats>

Quote:

Originally Posted by bigrigdriver

Try this:

Code:

find /test/upload/ -type f -name "*.xml" -exec grep -l -A 1 -B 1 "Sports" {} \; -print

-A 1 print one line After the line containing Sports
-B 1 print one line Before the line containing Sports

evo2 · 01-19-2010, 04:58 AM

Grep is usually used for single line matches. For multiline matching you may do better to look at sed.

Evo2.

guessity · 01-19-2010, 07:10 AM

Quote:

Originally Posted by evo2

Grep is usually used for single line matches. For multiline matching you may do better to look at sed.

Evo2.

thx. any idea how?

evo2 · 01-19-2010, 07:22 AM

There are oodles of sed tutorials on the web. Have a google (or wait for a sed expert to post the anser here).

Cheers,

Evo2

berbae · 01-19-2010, 11:31 AM

Code:

for file in $(find /test/upload/ -type f -name "*.xml" -print); do
    if $(sed -n '/<cats>/ {n
                           p}' $file|grep -q Sports); then
        echo $file
    fi
done

This seems to answer your request, if I understood what you want to get, ie only the name of the files with the pattern :
<cats> immediately followed on the next line by the word Sports.

schneidz · 01-19-2010, 12:25 PM

maybe egrep -n "(sports|cats)" to get the line number.
then
if expr `grep -n sports` - `grep -n cats` -eq 1
then print the filename.

but there must be a simpler way.

guessity · 01-20-2010, 04:57 AM

Quote:

Originally Posted by berbae

Code:

for file in $(find /test/upload/ -type f -name "*.xml" -print); do
    if $(sed -n '/<cats>/ {n
                           p}' $file|grep -q Sports); then
        echo $file
    fi
done

This seems to answer your request, if I understood what you want to get, ie only the name of the files with the pattern :
<cats> immediately followed on the next line by the word Sports.

works well. thx..