Shell Programming and Scripting

i need to sort to get all the unique records based on the 1st and 2nd column, and keep the record with the highest value on 5th column if there are duplicates, every column with varies length

a^2^x^y^z
bxc^2xx2^aa^bvxxxx^cdd
a^3^1^2^3
a^2^x^1^c

I want a result which will only keep the 1st 3 records.

thank you very much

---------- Post updated at 09:26 PM ---------- Previous update was at 08:03 PM ----------

if how to choose the dups are an issue, can just choose the 1st occurrence. get all the unique records based on the 1st and 2nd column is good enough.

thank you.

Try something like this:

Code:

$sort -t"^" -k1,1 -k2,2 -k5,5r file.txt | awk -F"^" '!a[$1,$2]++'
a^2^x^y^z
a^3^1^2^3
bxc^2xx2^aa^bvxxxx^cdd

thanks a lot.