Hi. I have two data files a.dat and b.dat. a.dat contains : s:b: h:g:wq and b.dat contains c:b d:b f:b w:r My output should be : s:b:c,d,f h:g:wq,r. Hpw can i accomplish this task in shell scripting Regards Arun

It's not pretty. Read a.dat and if it's line 1 grab fields 1, 2, 3 of file b and build the required string. if it's line 2 of a.dat, grab the 2nd field of line 4 of b.dat, and build the string: #!/bin/bash cnt=0 while read line do ((cnt+=1)) if [[ $cnt -eq 1 ]] then f1=$(awk ' BEGIN { FS=":" } { if (NR == 1) { print $1 ; exit } } ' b.dat) f2=$(awk ' BEGIN { FS=":" } { if (NR == 2) { print $1 ; exit } } ' b.dat) f3=$(awk ' BEGIN { FS=":" } { if (NR == 3) { print $1 ; exit } } ' b.dat) line="${line},${f1},${f2},${f3}" echo $line fi if [[ $cnt -eq 2 ]] then f4=$(awk ' BEGIN { FS=":" } { if (NR == 4) { print $2 ; exit } } ' b.dat) line="${line},${f4}." echo $line fi done < a.dat

Thanks for the message. But what i have shown is only a sample. What i meant was my first file 2nd field matches with 2nd file 2nd field. Delimiter is ":" (Colon). My final output should be 1 st file 1st field,1st file 2nd field which is comman to both and third field should be 2nd file 1st field. Regards Arun