topic Re: Missing Data in SAS Programming

Missing Data

Wafflecakes — Tue, 10 Oct 2017 14:35:00 GMT

To SAS experts - I have repeated measures data (multiple obs per id) and I am trying to duplicate a variable but only keep values that satisfy a condition:

Have

id x x1

1 1 2

1 2 .

1 2 30

2 1 350

2 2 .

2 2 30

Want

id x x1 x2

1 1 2 2

1 2 . .

1 2 30 .

2 1 350 350

2 2 . .

2 2 30 .

I tried this :

data try;

set try;

if x1~=. then x2=x1;

where x1=2;

run;

SAS deletes all of my other observations and only keeps those that follow the if statement - what do I do?

Re: Missing Data

Astounding — Tue, 10 Oct 2017 13:59:58 GMT

It sounds like you are looking for this:

data want;

set have;

if x1=2 then x2=2;

run;

Re: Missing Data

Wafflecakes — Tue, 10 Oct 2017 14:06:29 GMT

X1 equals values other than 2 like 350 so that wudnt work

Re: Missing Data

Astounding — Tue, 10 Oct 2017 14:12:48 GMT

You may need to explain a fuller set of rules then. Take the example in your data where X1 is 3. Why is X2 missing, and not 3?

Could the solution be as simple as this?

x2 = x1;

Re: Missing Data

Wafflecakes — Tue, 10 Oct 2017 14:21:25 GMT

I want x2 to be missing on purpose

Re: Missing Data

ballardw — Tue, 10 Oct 2017 14:24:43 GMT

The purpose of this statement is to remove any observation where the value of X1 is other than 2

where x1=2;

So you told SAS to delete observations.

Provide a description of the actual rule(s) involved.

Re: Missing Data

Astounding — Tue, 10 Oct 2017 14:25:35 GMT

OK, time for full disclosure then. No secrets allowed.

When should X2 be equal to X1?

When should X2 be missing?

Are there any other possibilities?

The programming is easy. You just have to spell out the rules.

Re: Missing Data

Wafflecakes — Tue, 10 Oct 2017 14:36:04 GMT

I modified the original post. Basically you will notice that x can equal 1 or 2. I only want to keep the values of x1 for when x=1 otherwise x2=. even if x1~=.

Re: Missing Data

Wafflecakes — Tue, 10 Oct 2017 14:40:51 GMT

I only added the where statement so that the if statement executes in those places and the remainder is meant to be missing.

Re: Missing Data

Astounding — Tue, 10 Oct 2017 14:47:11 GMT

Assuming I understand correctly, this would do the trick:

data want;

set have;

if x=1 then x2=x1;

run;

Re: Missing Data

Wafflecakes — Tue, 10 Oct 2017 14:50:37 GMT

This is it. Such a simple step i made it so overly complicated.

Re: Missing Data

ballardw — Tue, 10 Oct 2017 20:02:08 GMT

@Wafflecakes wrote:
I only added the where statement so that the if statement executes in those places and the remainder is meant to be missing.

Where does not work that way. Where subsets data given the rules on the statement. It actually works as data is read and does not matter where in the data step the where statement exists. Unless you have multiple where statements, in which case the last one is likely going to be the one in effect.