topic Re: Assigning Random numbers to a given set of IDs in SAS Programming

Assigning Random numbers to a given set of IDs

danwarags — Fri, 26 Aug 2016 14:56:43 GMT

Hello all,

I am trying to assign random numbers to a given set of IDs. The reason for assinging random numbers to the IDs is to de-identify the information from the IDs. There should not be any relationship between the IDs and the Random numbers generated from them. For example:

if I have a set of IDs as:

IDs	Random Numbers
1	5768
1	5768
2	2749
2	2749
2	2749
5	9574
5	9574
5	9574

The random numbers in the above tables are just the examples. These numbers can be any given number until and unless there is no relationship between the IDs and Random numbers. The only intention for decoding the IDs is to secure the information. Again, these random numbers cannot be used to lookback the IDs.

Thank you

Re: Assigning Random numbers to a given set of IDs

Ksharp — Fri, 26 Aug 2016 15:47:06 GMT


data have;
input id;
cards;
1
1
2
2
2
3
4
5
5
;
run;
data want;
 set have;
 by id;
 retain ran;
 if first.id then ran=ceil(1000000000000*ranuni(0));
run;

Re: Assigning Random numbers to a given set of IDs

danwarags — Thu, 01 Sep 2016 14:18:35 GMT

Thanks Ksharp. This code works. If I wanted to assign random numbers in particular range how should I modify the code. For example for a given set of 2000 IDs I want to assign random numbers in the range 100000- 102000. How should I do this. For example:

Obs	ID
1	1456
2	3765
3	35632
4	89564
.	.
.	.
.	.
2000	580058

My output should look something like:

Obs	ID	Random
1	1456	100000
2	3765	101997
3	35632	101456
4	89564	100002
.	.
.	.
.	.
2000	580058	102000

In the above example, the number of observations in the dataset are 2000. The Random numbers are assigned such that they are in particular range i.e 100000-102000.

Thanks for your help

Re: Assigning Random numbers to a given set of IDs

Ksharp — Fri, 02 Sep 2016 04:09:56 GMT



data have;
 do obs=1 to 2000;
  output;
 end;
run;
data temp;
 set have;
 ran=ranuni(0);
run;
proc rank data=temp out=want;
 var ran;
 ranks r;
run;
data want;
 set want;
 random=100001+r;
 drop r;
run;

Re: Assigning Random numbers to a given set of IDs

Ksharp — Fri, 02 Sep 2016 04:12:31 GMT

Last code is not right. Try this one:



data have;
 do obs=1 to 2000;
  output;
 end;
run;
data temp;
 set have;
 ran=ranuni(0);
run;
proc rank data=temp out=want;
 var ran;
 ranks r;
run;
data want;
 set want;
 random=100000+r;
 drop r;
run;

proc sql;
select count(distinct random),max(random),min(random) from want;quit;