[SOLVED] find script

shellarchive · 08-18-2010, 02:07 AM

i am trying to write scripts to pass 2 different variables to find command.

find . -name $var1 -exec grep -H $var2

for example: find . -name CDR_2010-07-21 -exec grep -H 9892614477 {} \;
1st variable CDR_2010-07-21(passed by var1) is the file name inside which i am trying to search string 9892614477(passed by var2) .

i have tried following script

" for file in CDR_2010-07-20* CDR_2010-07-12* CDR_2010-08-05* CDR_2010-08-14* CDR_2010-08-11* ;
do
for file2 in 9990035233 9718290475 9899867969 9820048711 9830589910 9891446365;
do find /data6/ -name $file -exec grep -H $file2 {} \; ;
done;
done; "

but the problem is that it searches the each string in all the CDR_ files.

druuna · 08-18-2010, 02:17 AM

Hi,

Why are you using this: for file in CDR_2010-07-20* CDR_2010-07-12*......

Bash will expand CDR_2010-07-20* to all files that match. Example: if CDR_2010-07-20_1 and CDR_2010-07-20_2 exist, these 2 will be substituted for CDR_2010-07-20*.

The below code will search for each number in file2 in every CDR file mentioned:

Code:

for file in CDR_2010-07-20 CDR_2010-07-12 CDR_2010-08-05 CDR_2010-08-14 CDR_2010-08-11
do
  for file2 in 9990035233 9718290475 9899867969 9820048711 9830589910 9891446365
  do
    find /data6/ -name $file -exec grep -H $file2 {} \;
  done
done

Hope this helps.

shellarchive · 08-18-2010, 02:41 AM

i really appreciate your quick reply.

however what i want is to search first number say 9990035233 in first cdr file ie CDR_2010-07-20.

in simpler words, script should execute command like this,

find /data6/ -name CDR_2010-07-20 -exec grep -H 9990035233 {} \;
find /data6/ -name CDR_2010-07-12 -exec grep -H 9718290475 {} \;
find /data6/ -name CDR_2010-08-05 -exec grep -H 9899867969 {} \;

and so on.

druuna · 08-18-2010, 03:06 AM

Hi,

Your example data is not consistent: There are 5 files and 6 search terms.

This should work if both are equal (I added a dummy file):

Code:

#!/bin/bash

# put data in 2 arrays
files=(CDR_2010-07-20 CDR_2010-07-12 CDR_2010-08-05 CDR_2010-08-14 CDR_2010-08-11 CDR_2000-01-01)
sterms=(9990035233 9718290475 9899867969 9820048711 9830589910 9891446365)

# show content arrays (can be removed, just here to show content)
echo ${files[@]}
echo ${sterms[@]}

# loop over all elements of arrays
for (( x=0 ; x<=${#files[@]} ; x++ ))
do
  # use element number for specific content
  find /data6 -name "${files[$x]}" -exec grep -H "${sterms[$x]}" {} \;
done

Hope this helps.

crts · 08-18-2010, 03:07 AM

Quote:

Originally Posted by shellarchive

i really appreciate your quick reply.

however what i want is to search first number say 9990035233 in first cdr file ie CDR_2010-07-20.

in simpler words, script should execute command like this,

find /data6/ -name CDR_2010-07-20 -exec grep -H 9990035233 {} \;
find /data6/ -name CDR_2010-07-12 -exec grep -H 9718290475 {} \;
find /data6/ -name CDR_2010-08-05 -exec grep -H 9899867969 {} \;

and so on.

Hi,

you already have the solution for your problem. The for loops will search, as you already observed, for every grep pattern in every file.
To match all files that start with CDR_xxx you can, e. g.

Code:

find /data6/ -name 'CDR_2010-07-20*' -exec grep -H 9990035233 {} \;
find /data6/ -name 'CDR_2010-07-12*' -exec grep -H 9718290475 {} \;
find /data6/ -name 'CDR_2010-08-05*' -exec grep -H 9899867969 {} \;

Or do you want a script, where you pass the files and patterns as parameters? E. g. something like

./search.sh file1 pattern1 [file2 pattern]

Code:

#!/bin/bash

while [[ $# -ge 2 ]]; do
    find /data6/ -name "$1" -exec grep -H "$2" {} \;
    shift
    shift
done

(UNTESTED)

shellarchive · 08-18-2010, 03:57 AM

thanks druuna & crts for quick solution.

both the script did the work.

but i found "for loop" script (posted by druna) usefull & real because i have to pass 300 to 400 such arguments.

one thing which i didn't understand in your script was "x<=${#files[@]}" in for statement. would appreciate if you elaborate this.

i was trying it in another way
by putting all CDR_ files name inside textfile a.
"[root@test /tmp]cat a
CDR_2010-07-20
CDR_2010-07-12
CDR_2010-08-05
CDR_2010-08-14
CDR_2010-08-11
[root@test /tmp]"

and then using xarg & echo command appending the CDRnames from file a to echo find . -name .

"[root@test /tmp]cat a|xargs -n1 echo "find /data6/ -name "
find /data6/ -name CDR_2010-07-20
find /data6/ -name CDR_2010-07-12
find /data6/ -name CDR_2010-08-05
find /data6/ -name CDR_2010-08-14
find /data6/ -name CDR_2010-08-11
"

but after this i was finding it impossible to append "-exec grep -H 9718290475" to the existing output.

druuna · 08-18-2010, 04:07 AM

Hi,

This: x=0 ; x<=${#files[@]} ; x++ is a counter that starts with 0 (x=0) and keeps adding 1 (x++) until the amount of elements in the array is reached (x<=${#files[@]}).

${#files[@]} returns the number of elements in an array called files.

Have a look here: Advanced Bash-Scripting Guide - Chapter 27. Arrays for more details on arrays in bash.

Hope this clears things up a bit.

BTW: You're welcome

crts · 08-18-2010, 04:31 AM

Quote:

Originally Posted by shellarchive

i was trying it in another way
by putting all CDR_ files name inside textfile a.
"[root@test /tmp]cat a
CDR_2010-07-20
CDR_2010-07-12
CDR_2010-08-05
CDR_2010-08-14
CDR_2010-08-11
[root@test /tmp]"

and then using xarg & echo command appending the CDRnames from file a to echo find . -name .

"[root@test /tmp]cat a|xargs -n1 echo "find /data6/ -name "
find /data6/ -name CDR_2010-07-20
find /data6/ -name CDR_2010-07-12
find /data6/ -name CDR_2010-08-05
find /data6/ -name CDR_2010-08-14
find /data6/ -name CDR_2010-08-11
"

but after this i was finding it impossible to append "-exec grep -H 9718290475" to the existing output.

Hi,

with your cat|xargs approach you will face the same problem that every file will be searched for the search pattern.
So regard this just as complementary info on xargs.

Try the following

Code:

cat a | xargs -n1 -I [] echo "find /data6/ -name [] -exec grep 'pattern' {} \;"

konsolebox · 08-18-2010, 05:43 AM

you can put everything in pairs like this one

input file:

Code:

CDR_2010-07-20:9990035233
CDR_2010-07-12:9718290475
CDR_2010-08-05:9899867969
CDR_2010-08-14:9820048711
CDR_2010-08-11:9830589910

script:

Code:

#!/bin/bash
while read; do
    A=${REPLY%:*} B=${REPLY#*:}
    find /data6/ -name "$A" -exec grep -H "$B" {} \;
done

run:

Code:

bash script.sh < input_file.list

Same concept can be made if you want to add custom target directories instead of just data6.

konsolebox · 08-18-2010, 05:50 AM

Or in separate files:

Code:

#!/bin/bash

LIST1=$1 LIST2=$2

if [[ -f $LIST1 && -f $LIST2 ]]; then
    while read -u 3 A && read -u 4 B; do
        find /data6/ -name "$A" -exec grep -H "$B" {} \;
    done 3<"$LIST1" 4<"$LIST2"
fi

Code:

bash script.sh input_file_a.list input_file_b.list

Same concept can be applied for 3 file list.