topic Re: Grouping by flag in SAS Programming

Grouping by flag

RandyStan — Thu, 13 Feb 2020 05:11:35 GMT

Dear All:

I want to create 10 groups for the variable VARA using the 0/1 flag. So my data is as follows

VarA Flag

10 0

8 0

11 1

24 1

35 0

and so forth

I want to group the data by VARA and the flag.

So I want

VarA Flag Group_FLAG_0 Group_FLAG_1

10 0

8 0

11 1

24 1

35 0

Can this be done in one step.

Thanks in advance

Randy

Re: Grouping by flag

andreas_lds — Thu, 13 Feb 2020 06:06:45 GMT

Maybe it can be done in one step, but since you have not revealed the logic to be applied in an (for me) understandable way, i can't give you any details on how to do it. And, as side-note, you should start posting data in usable form, so that there is one thing less we need to guess.

Re: Grouping by flag

ballardw — Thu, 13 Feb 2020 15:25:26 GMT

What values would be assigned to Group_FLAG_0 and Group_FLAG_1?

What are the rules for assigning the values to the variables?

Re: Grouping by flag

Krueger — Thu, 13 Feb 2020 15:57:19 GMT

How do you want them grouped? What's the logic?

data have;
	input VarA Flag;
	datalines;
10 0
8  0
11 1
24 1
35 0
	;
run;
proc sort data = have; by VarA; run;
data want;
	set have;
	by varA;
	if Flag = 0 then Group_FLAG_0 = 1;
	else if Flag = 1 then Group_FLAG_1 = 1; 
run;

Re: Grouping by flag

RandyStan — Thu, 13 Feb 2020 21:04:57 GMT

If I may, let me try to explain

I have varA with numbers and a Flag (0,1)

I want to construct a decile column (which gives the decile for VarA when Flag = 0; and when Flag = 1

So for example; (A shortened version of the data set)

VarA Flag

10 0

12 0

3 1

15 1

25 0

And I construct a group_Rank variable such that

VarA Flag GroupRank_VARA

10 0 1

12 0 2

15 1 2

3 1 1

25 0 3

proc rank data=have out= want groups=10;

by Flag;

var VarA;

rank GroupRank_VarA;

run;

I hope I am clearer this time.

Randy

Re: Grouping by flag

ballardw — Thu, 13 Feb 2020 23:47:22 GMT

I think if you display your want in a clearer form:

VarA       Flag      GroupRank_VARA
10           0                1
12           0                2
25           0                3
 3           1                1
15           1                2

You want:

proc rank data=have out= want groups=10; 
   by Flag; 
   var VarA; 
   ranks GroupRank_VarA; 

run;

Note the S in RANKS.

Your code would have been throwing errors that maybe you should start with.

And as with all BY statements you have to sort the data by the variables on the BY statement before use.

Note that Proc Rank will create ranks of 0 to 9 when requesting groups=10. So if you really must have 1 to 10 you will need to take a pass through a data step to add 1.

From the documentation for Proc Rank

GROUPS=number-of-groups

assigns group values ranging from 0 to number-of-groups minus 1.