Shell Programming and Scripting

paste file1 file2 # Here we join files line by line

awk -F "[,\t]" # here we have comma separated file and after paste it has tab which separates two files.

for(i=1;i<=(NF/2);i++){if($i != $(NF/2+i)) # we have to compare column 1 of first file to column 1 of second file and so on. So i have divided number fields to compare it. 1 to (NF/2)+1 and so on..

if(s){s=s";"$i}else{s=$i}}else{if(s){s=s";,"}else{s=","}}}}{ print s;s=""} # Here just we assign values to the variable s as per your requirement and at the end we print that s.

Hope this helps you

my file has around 1 lakh rows and 300 columns.

Though the command mentioned above is working fine but is taking more than 2 hrs to completely execute and is time consuming.

Is there any command that can specify only the column values and column number where we have the difference like

If there is difference at column number 3 and column number 7, resultant file will have value as
3,7 (Column Number)
Rahul,23 (Values present in file 1)

Rather than displaying "," for all the columns that are matching between two files. It only display columns where we have the mismatch but with a column number to keep the track.

Thanks

I am not sure how we can minimize time for this...

as per above requested input..

is this what you want..?

Code:

$paste file1 file2 | awk -F "[,\t]" '{for(i=1;i<=(NF/2);i++){
if($i != $(NF/2+i)){if(s){s=s";"$i","$(NF/2+i)"("i")"}else{s=$i","$(NF/2+i)"("i")"}}}}{ print s;s=""}'

 16-12-2011, 14-12-2011(2)
 14-12-2011,18-12-2011(2)
Maik,Malik(1);113,11(3)

Thanks for the Help

Its woking exactly the way I want to. Only thing is it's taking much time to execute completely. Don't know how to overcome that