Great piece of code Ygor !!

Little modification to yr script to handle line numbers in the file.

awk '{FS=":"}{x="";x=tolower($1)+tolower($2);a[x]++;b[x]=$0}END{for(x in a)if(a[x]==1)print b[x]}' file1

question:

I have about 150 files to process every night.
The size of each file can be between 5Mb to 20Mb.

What is more efficient to use?

thanks you all guys, you are great and I like this forum very much.

Thank for your patient.

I am java programmer but sometimes I have some unix stuff to do which is not my cup of tea... But I begin to like it because of you

The awk solution is a clear winner in the performance department. It is only one process. The other solution has one grep process per line and a handful of other processes.

I find this very helpful and hope that other people like me can use it on the future.