topic Creating new variable in New SAS User

Creating new variable

StudyBeane — Mon, 09 Sep 2024 10:11:21 GMT

Hi, here is my predicament. I'm taking a beginner SAS course concurrently with another course that relies on the information you learned in that beginner class. I have ZERO SAS experience, only R. One of the questions is like this:

"Create a new variable called “abp” that is coded 1 if the individual’s daily alcohol consumption was 80 g per day or higher, and that is coded 0 if the individual’s daily alcohol consumption was less than 80 g per day."

In the dataset, the acgp (original variable) is:

0 = 0-39 grams per day

1 = 40-79 grams per day

2 = 80-119 grams per day

3 = 120+ grams per day

So, just by scouring around youtube and other resources, I did:

if acgp in (2,3) THEN abp = 1;
ELSE IF acgp in (0,1) THEN abp = 0;
run;

Now this will run but the output data is nothing, it gives me two columns with acgp and acp and no rows. The code I received was from a few basic codes given to me. I don't know SAS at all, so I don't know if I have to do anything else before doing the IF/ELSE statement? My data is a SAS data set, its already read into my program. Any advice would be VERY much appreciated!

Re: Creating new variable

PaigeMiller — Mon, 09 Sep 2024 11:22:09 GMT

Show us the entire data step code, not a few selected lines.

Show us the log from running this data step (we need to see the code from this data step as it appears in the log, and any ERRORs/WARNINGs/NOTEs that appear in the log)

Re: Creating new variable

Kurt_Bremser — Mon, 09 Sep 2024 11:30:48 GMT

If you did not use a SET statement to read the existing dataset, you only get these two variables with missing values in a single observation.

Please share the complete code of the step (from the DATA to the RUN statement).

Re: Creating new variable

Tom — Mon, 09 Sep 2024 12:03:35 GMT

You left out two key pieces of information needed to create a program to do what you wanted.

The NAME of the existing dataset that has the variable ACGP.

The NAME of the NEW dataset you want to create that will have the NEW variable ABP.

Let's assume those are HAVE and WANT, respectively.

Also your logic does not have any instructions for what to do when the value ACGP is something else, such as a missing value or an invalid code.

data want;
  set have;
  if acgp in (2,3) then abp = 1;
  else if acgp in (0,1) then abp = 0;
  else abp=.;
run;

If you are willing to have missing or invalid codes mapped to 0 then you can simplify the logic to just assigning the result of the first boolean test to ABP.

data want;
  set have;
  abp = (acgp in (2,3));
run;

That will work because SAS evaluate boolean expressions to either 1 (TRUE) or 0 (FALSE).

Re: Creating new variable

StudyBeane — Mon, 09 Sep 2024 13:48:51 GMT

Yes, thank you, I am extremely new to SAS and didn't realize I had to use use data/set.

Re: Creating new variable

ballardw — Mon, 09 Sep 2024 15:04:40 GMT

@StudyBeane wrote:

Yes, thank you, I am extremely new to SAS and didn't realize I had to use use data/set.

In R don't you have to provide some data source before creating a new variable based on existing values?

Pretty much the case in any of the 20+ "programming languages" I have used over the years. No input data = no meaningful output data. The principles wont change but approach will change based on the language used. Such as how to point to the existing data: Set statement for a SAS data set or DATA= for almost every other SAS procedure, and where to put the output: Data statement with the Data step code. Other approaches in other procedures but OUTPUT or OUT= are fairly common but the syntax can vary based on the procedure.

SAS will default to the last created data set as input for many procedures but the Data step still needs a SET statement to do such even if no data set name is provided.