Shell Programming and Scripting

Hello all,
I have a input file like this.

input file
---------------
abc ab001 + ab002 zca
acb ab006 + ab007 caz
cba ab003 + ab004 zca
bac ab004 - ab005 zac
bca ab002 - ab003 cza
cba ab005 + ab006 acz
cba ab005 + ab006 acz
cab ab007 + ab008 azc
bca ab002 + ab003 cza
cba ab003 + ab004 zca
bac ab004 + ab005 zac
acb ab006 - ab007 caz
abc ab001 - ab002 zca
cab ab007 - ab008 azc

desired output file
--------------------------
col 1 col2
---- ----
3 3
5 -2
11 6
14 -3

Explanation :-
Actually i want to print continuous "+" pattern present in column number 3 then print its occurrence number then "-" patterns occurrence number (col 2).
while (col 1) is line number in which last pattern was present.
In another way it just absolute addition (without considering sign) of values in (col 2).

Sorry for my english.
please help me out

Code:

awk 'END { print NR, s c }
NR > 1 && $3 != _3 { print NR - 1, s c; c = 0 }
{ _3 = $3; c++; s = /+/ ? x : $3 }' infile

Use gawk, nawk or /usr/xpg4/bin/awk on Solaris.

Can any one explain this code ?

awk 'END { print NR, s c } # Give the last output
NR > 1 && $3 != _3 { print NR - 1, s c; c = 0 } # When line >1, $3 is changed (compare with previous line) print the result. after print, reset count to zero.
{ _3 = $3; c++; s = /+/ ? x : $3 }' infile # reserve current $3 value, count the continuous time, reserve the - pattern status. x is nothing,empty

Change to below code, will be easy understand and readable.

Code:

 awk ' NR > 1 && $3 != temp { print NR - 1, s c; c = 0 }
{ temp = $3; c++; s = /+/ ? x : $3 } 
END { print NR, s c }'

below perl may help you some

Code:

my $key;
while(<DATA>){
	my @tmp = split;
	if($tmp[2] ne $key){
		if($key ne ""){
			print $.-1," (",$key,")",$cnt,"\n";
			$key=$tmp[2];
			$cnt=0;
		}
		else{
			$key=$tmp[2];
		}
	}
	$cnt++;
}
print $.," (",$key,")",$cnt,"\n"; 
__DATA__
abc ab001 + ab002 zca
acb ab006 + ab007 caz
cba ab003 + ab004 zca
bac ab004 - ab005 zac
bca ab002 - ab003 cza
cba ab005 + ab006 acz
cba ab005 + ab006 acz
cab ab007 + ab008 azc
bca ab002 * ab003 cza
cba ab003 * ab004 zca
bac ab004 + ab005 zac
bca ab002 + ab003 cza
cba ab003 + ab004 zca
bac ab004 + ab005 zac
acb ab006 - ab007 caz
abc ab001 - ab002 zca
cab ab007 - ab008 azc

Another one in awk:

Code:

awk '/+/ { ++p; if(n)print NR-1"-"n;n=0; } /-/ { ++n; if(p)print NR-1" "p;p=0 } 
         END{ if(p) print NR" "p+n; else print NR"-"p+n; }' file