Counting unique values by subsets of a variable

[AJChamberlain]

Greetings community

I am trying to work out a way to sum the number of unique values that occur for a variable in the dataset. So for example if the values for my observations were:

1

1

2

2

3

4

6

The answer would be 5 because there are 5 unique values. But I want to find this answer for each combination of two other variables in my dataset (type and version), so my data looks like this, where I am trying to count the number of unique instances of 'value':

Value	Type	Version
1	1	A
1	1	A
2	1	A
2	1	A
3	1	A
4	1	A
6	1	A
1	2	A
1	2	A
3	2	A
3	2	A
1	1	B
3	1	B
4	1	B
5	1	B
1	2	B
2	2	B
3	2	B
4	2	B
5	2	B
6	2	B

and my answer needs to look something like this:

	Version
Type	A	B
1	5	4
2	2	6

I think I can do this with Proc SQL but it has eluded me so far, any suggestions gratefully received!

Thanks

Andy

[PaigeMiller · Accepted solution]

data a;
	input value type version $;
	cards;
1	1	A
1	1	A
2	1	A
2	1	A
3	1	A
4	1	A
6	1	A
1	2	A
1	2	A
3	2	A
3	2	A
1	1	B
3	1	B
4	1	B
5	1	B
1	2	B
2	2	B
3	2	B
4	2	B
5	2	B
6	2	B
;
run;

proc sql;
	select type,version,count(distinct value) from a group by type,version;
quit;

Now this does the calculations of the number of distinct levels of value, it does not re-arrange the results into a table with two rows (one for each type) like you show, but that's pretty simple to do in PROC REPORT if you really have to have it in that form.

[AJChamberlain]

Thakns Paige, that's worked for me. A

[novinosrin]

Your sample output doesn't look like a dataset rather report with proc report/tabulate