Splitting a file based on two patterns

Shell Programming and Scripting
1

Hi there,

I've an input file as follows:

*START
1001 a1
1002 a2
1003 a3
1004 a4
*END
*START
1001 b1
1002 b2
1004 b4
*END
*START
1001 c1
1004 c4
*END

I would like to split this file into three files such that the contents of first file contains all the rows between the first set of *START and *END, the contents of the second file contains the rows between the second set of *START and *END and so on.

My output will be three files:

file1
1001 a1
1002 a2
1003 a3
1004 a4

file2
1001 b1
1002 b2
1004 b4

file3
1001 c1
1004 c4

Any assistance is greatly appreciated.

2 
$
$ cat kbirde.txt
*START
1001 a1
1002 a2
1003 a3
1004 a4
*END
*START
1001 b1
1002 b2
1004 b4
*END
*START
1001 c1
1004 c4
*END
$
$ perl -ne 'BEGIN {$/=""; $i=1;}
> {
>   while (/\*START.(.*?)\*END/gs) {
>     open (F,">file".$i++); print F $1; close(F);
>   }
> }' kbirde.txt
$
$ cat file1
1001 a1
1002 a2
1003 a3
1004 a4
$
$ cat file2
1001 b1
1002 b2
1004 b4
$
$ cat file3
1001 c1
1004 c4
$

Hope that helps,
tyler_durden

______________________________________________
"Only after disaster can we be resurrected."

3

Hi Tyler,

Thanks a lot. Can this be written in awk (bash shell)?

Regards,
kbirde

4

Yes, sure,you can definitely use awk to do it as well.

Try this. This should create as many files as there are START and END pair

awk 'BEGIN{i=1}/\*START/{getline;n=0
while ($0 !~ /\*END/){s=(n==0) ? $0: s"\n"$0;n=i;getline}print s > "file"'"i"';
print "one over";i++
}' filename

cheers,
Devaraj Takhellambam

5

Hello Devaraj,

Thanks so much for the assistance.

Cheers,
kbirde

6

Another approach:

awk '/START/{p=1;i++;next}
/END/{p=0;next}
p{print > "file" i}' file
7

perl:

$/="*END\n";
open $fh,"<","a.txt";
while(<$fh>){
	open FH,">file$..txt";
	print FH $_;
	close FH;
}
close $fh;

awk:

nawk 'BEGIN{n=1}
{
file=sprintf("%s.txt",n)
print $0 >> file
if ($0 ~ /^*END/)
n++
}
' a.txt