Dear deepakiniimt, I'm a little confused, so please forgive me if I have missed the point. Perhaps putting code tags in your message would help clarity. It seems you may have chopped a bit out too, or i just don't "get it"

You say you have a bunch of files that you expect to be the same apart from a few differences. Could you consider the logic that if one file is the standard, then you could diff each one against it? Something like:-

Code:

for file in `ls -1 $dir/$files`
do
   echo $file
   diff standard.file $file
done > outputfile

Would that thinking suffice?

If you are looking for records that are unique in your list of files, then perhaps you should think about:You could:-

Code:

cat $dir/$files|uniq -u>/tmp/unique_records
grep -f /tmp/unique_records $dir/$files

Depending on your version of unix, you may need fgrep instead.

If this matches too many lines, then it may be that the lines are sometimes part of larger lines, eg:-

Code:

cat /tmp/unique_records
Hello said dad
Hello said mum
Hello Hello Hello
Hello

In this case, the last record would match the other three, along with any other record in any file that had the simple string "Hello". You could edit the file of unique records to make them all start with a carat symbol ^ and end with a dollar $ to explicitly signify begginning and end of line and the two ends of your string. You code would grow to:-

Code:

cat $dir/$files|uniq -u|while read line
do
   echo "\^${line}\$"
done >/tmp/unique_records
grep -f /tmp/unique_records $dir/$files

It's a bit clumsy and perhaps a sed or an awk would be quicker, just depending how much data you have.


Does that meet your needs, or have I gone off track completely?

I hope that it helps,
Robin
Liverpool/Blackburn
UK