UNIX for Advanced & Expert Users

Hi Gurus,

I have a file

1|usa|hh
2|usa|ll
3|usa|vg
4|usa|vg
5|usa|vg
6|usa|vg
7|usa|ll
8|uk|nn
9|uk|bb
10|uk|bb
11|kuwait|mm
12|kuwait|jkj
13|kuwait|mm
14|dubai|hh

I want to group by last two columns and get the last two recs and count.

output should look like :

dubai hh 1
kuwait jkj 1
kuwait mm 2
uk bb 2
uk nn 1
usa hh 1
usa ll 2
usa vg 4

I tried something like this but this doesnt give the right counts

awk -F"|" '{col[$2,$3]=NR} END {for (i in col) print i, col[i]}' new.txt | sort

dubaihh 14
kuwaitjkj 12
kuwaitmm 13
ukbb 10
uknn 8
usahh 1
usall 7
usavg 6

Can any of the Unix Gurus help.
It has been posted earlier but I have modified it a bit.
Any suggestion is very much appreciated.


Thanks
Sumeet

Code:

awk -F"|" '{col[$2,$3]++} END {for (i in col) print i, col[i]}' new.txt | sort

Thanks Jim

It works perfectly but one qusn : I couldnt understand the functionality of

{col[$2,$3]++} in comparison with {col[$2,$3]=NR}.

Can you throw some light on it.

Thanks
Sumeet

NR is the current Number of record, in other words the line number of the file. When adding this you are not counting the occurrences but the positions in the file