topic Re: Using first. to add a count unique by two variables in SAS Programming

Using first. to add a count unique by two variables

angel2 — Wed, 25 Nov 2020 16:28:34 GMT

How can I use a data step to create a count unique by ID and category. I want the same count to remain for non unique values. Please see example attached and the wanted counts. Thanks

Re: Using first. to add a count unique by two variables

novinosrin — Wed, 25 Nov 2020 16:39:35 GMT


data have;
input (ID	CATEGORY) ($);*		WANTED COUNT;
cards;
001	PIG		1
001	PIG		1
003	SHEEP		1
003	PIG		2
003	PIG		2
004	SHEEP		1
004	SHEEP		1
005	PIG		1
005	PIG		1
005	COW		2
005	SHEEP		3
;

data want;
 set have;
 by id category notsorted;
 if first.id then count=1;
 else if first.category then count+1;
run;

Re: Using first. to add a count unique by two variables

RichardDeVen — Wed, 25 Nov 2020 16:44:48 GMT

Presuming the data is grouped into contiguous blocks the syntax BY var1 var2 NOTSORTED; can be used. This assumption is made because group 5 is not sorted in alphabetical order.

data have; input ID CATEGORY $ wanted; datalines; 001 PIG 1 001 PIG 1 003 SHEEP 1 003 PIG 2 003 PIG 2 004 SHEEP 1 004 SHEEP 1 005 PIG 1 005 PIG 1 005 COW 2 005 SHEEP 3 ; data want; set have; by id category notsorted; if first.id then seq=0; if first.category then seq+1; run;

Output

Re: Using first. to add a count unique by two variables

angel2 — Wed, 25 Nov 2020 17:02:19 GMT

Thank you so much - that's perfect!