topic Re: Percentile Buckets using Proc Rank in SAS Procedures

Percentile Buckets using Proc Rank

jacob_klimek — Tue, 01 Nov 2016 14:07:11 GMT

Hi,

I am trying to create buckets (bins) based off of the deciles of a variable. I.e I want the first 10% to receive a 1, the 10-20% to receive a 2, etc.

Below is an example of what my data looks like:

ID Var

1 2

2 25

3 12

4 16

5 8

6 13

7 9

8 11

9 3

10 1

And I want:

ID Var Bucket

1 2 2

2 25 10

3 12 7

4 16 9

5 8 4

6 13 8

7 9 5

8 11 6

9 3 3

10 1 1

I have been trying to work through the syntax using proc rank because I think that's what I want but I just can't get it. I think I am getting frustrated too easily because when I do a similar procedure in Stata it is pretty simple to grasp.

Any help would be great, thank you.

Re: Percentile Buckets using Proc Rank

RW9 — Tue, 01 Nov 2016 14:12:20 GMT

Would not proc surveyselect be a better option:

https://support.sas.com/documentation/cdl/en/statug/63347/HTML/default/viewer.htm#statug_surveyselect_sect001.htm

Re: Percentile Buckets using Proc Rank

ballardw — Tue, 01 Nov 2016 14:17:27 GMT

The basic bit is going to look like this:

proc rank data=have out=want groups=10;
   var var;
   ranks bucket;
run;

However proc rank assigns 0 to the lowest rank so you get 0 to 9 as provided in the documentation.

So you put your output data set through a data step to add 1;

data want;

set want;

bucket = bucket+1;

run;

Re: Percentile Buckets using Proc Rank

RW9 — Tue, 01 Nov 2016 14:21:15 GMT

Just to add an example of surveyselect:

data have;
  do id=1 to 10;
    var=4;
    output;
  end;
run;
  
proc surveyselect data=have groups=10 seed=49201 out=RandomGroups noprint;
run;