finding lines only between a certain string

Shell Programming and Scripting
1

Dear experts,

Ive been trying to figure this out for a while, but i cant. Please help.
I have a file, with approx 1 million lines. The contents are separated with "----------". Please see example below

So my problem is, i need to find all texts that have the keyword "GAA", but i need to see everything that is between the --------- lines. Desperatedly need a script to do this. Any help is appreciated. Thanks !

2

Can you provide sample expected output.

3

A sample expected output should be all lines in the file, which match the search pattern "GAA", and output in the format below

4

Try this:

awk '
        # new record, reset array index
        /^----/ { i=0 }
        # accumulate record contents in array
        { a[++i]=$0 }
        # matching record
        /GAA/ {
                # dump array contents
                for (j=1; j<=i; j++)
                        print a[j]
                # get the rest of this record
                while (getline && $0 !~ /^----/)
                        print
                # print the record terminator and reset the array
                print
                i=0
        }
' inputfile > outputfile
5

Annihilannic,

Tried the script, it took about 3 mins to run, but no output. outputfile was empty :frowning:

6

Strange, works fine for me with the sample data you provided. What operating system are you using? Are there any spaces at the beginning of the lines?

7

The file starts like this

Im using solaris 9

8

What's the output of head -50 /tmp/AIALARM_MSNLA_20080626.txt | cat -vet? Can you post it between [ code ] tags instead of [ quote ] tags please?

9

oh dear im getting a wierd result. the file is filled with ^M. Im sorry ill ftp the file out again. Thanks

root@ckpgpay11core> head -50 AIALARM_MSNLA_20080626.txt | cat -vet
Mroot@ckpgpay11core>
10

No big deal, it just means the file is in DOS/Windows format. You can use dos2unix to convert it (if available on your system), or tr -d '\r' < dosfile > unixfile.

11

Thanks guys

12

Dear Annihillanic,

I got the script working. Sorry but i have another question. I am making the script s little interactive, and have modified it to search for different strings everytime i run it. I cant seem to replace the variable "GAA" in the script. Any idea why? below if the output of the script when run in debug mode

root@ckpgpay11core> ./sarascript.sh
+ echo Enter input filename:
Enter input filename:
+ read inputfile
mstgb_aialarm_20081021.wri
+ echo Enter path of outputfile with name:
Enter path of outputfile with name:
+ read outputfile
/tmp/testing
+ echo Enter tag to look for:
Enter tag to look for:
+ read GAA
20081013003507
+ awk
        # new record, reset array index
        /^----/ { i=0 }
        # accumulate record contents in array
        { a[++i]=$0 }
        # matching record
        /\$GAA/ {
                # dump array contents
                for (j=1; j<=i; j++)
                        print a[j]
                # get the rest of this record
                while (getline && $0 !~ /^----/)
                        print
                # print the record terminator and reset the array
                print
                i=0
        }
 mstgb_aialarm_20081021.wri

root@ckpgpay11core> ls -la | grep testing
-rw-r--r--   1 root     other          0 Oct 30 10:47 testing
13

perl

undef $/;
open FH,"<file";
$str=<FH>;
@arr=split(/-+/,$str);
for $key(@arr){
if($key=~/pat/){
print $key;
}
}
close FH;
14

Because the awk script is between single quotes, shell variables are not expanded.

You can change it like this to allow it to interpolate the variable:

        /'"$GAA"'/ {

Or else use :

awk -v SEARCHSTRING="$GAA" '
        ...
        $0 ~ SEARCHSTRING {
        ...
' inputfile > outputfile
15

thanks guys. works perfectly