Data step counting by 2 variables

[MattehWoo]

Hi,

I have a code to count duplicated items:

data work.rank;

set work.Inbound;

by ITEM;
if first.ITEM then Rank=1;
else Rank+1;
Run;

This is great and does the job i want.

However, i'm now bringing through data from different countries so i now have a column that shows which country the items are in. In order for this code to work i have to have the data sorted by Item so that it can count. However i don't want to include different country data in the count. So if there is 1 item in GB 5 times it will rank it 1,2,3,4 and 5 etc, but when it goes to say FR and has the same Item, i want that one to start at 1.

eg:

Country	Item	Rank
GB	123	1
GB	123	2
GB	123	3
GB	1234	1
GB	1234	2
FR	12	1
FR	123	1
FR	123	2
FR	1234	1

The original code i had was perfect for use in 1 country but as soon as i brought through more it is failing as i tried sorting by country first and then item.

any help would be brilliant.

Apologies if what i am asking is silly.

Cheers,

M.

[Kurt_Bremser · Accepted solution]

Sort by item and country, and then use

by item country;

and first.country for the condition in your datastep.

Or simply sort by country and item, and use

by country item;

in your original datastep with the same first. condition.

[RW9]

Why not use the proc rank function, its what its made for:

http://support.sas.com/documentation/cdl/en/proc/61895/HTML/default/viewer.htm#a002473611.htm