topic how do you combine two variables into one variable in the same dataset in New SAS User

how do you combine two variables into one variable in the same dataset

whyamihere — Fri, 21 Feb 2025 00:34:25 GMT

I have been working on this code for three hours for a school assignment, can you please help? I have a large dataset where I need to combine the ethnicity and race and create a variable called group with the following values: 1 (black non-Hispanic) , 2 (white non-Hispanic), and 3 (Hispanic).

The following is the "key"

"1"= "American Indian or Alaska Native"
               "2"= "Asian"
               "3"= "Black or African American"
               "4"= "Native Hawaiian or Other Pacific Islander"
               "5"= "White"
               "6"= "Other"
               "7"= "Unknown";

"E1"= "Hispanic or Latino"
                    "E2"= "Non-Hispanic or Latino"
                    "E7"= "Unknown";

I think I might be on to something here with this code but I think I need to make a variable named "group"

/*Create a new variable called group*/
DATA=APPENDICITIS.PROJECT3;
if ETHNICITY = "E2" AND RACE = "3" THEN GROUP = "1";
if ETHNICITY = "E2" AND RACE = "5" THEN GROUP = "2";
IF ETHNICITY = "E1" THEN GROUP = "3";
RUN;

Thanks so much for your expertise and time.

Re: how do you combine two variables into one variable in the same dataset

SASKiwi — Fri, 21 Feb 2025 00:56:35 GMT

There's no equals sign in a DATA statement and LIBREFs are limited to 8 characters so I've shortened APPENDICITIS to APP. This logic is a bit better:

DATA APP.PROJECT3;
  set ???; * Need to read your input dataset;
if ETHNICITY = "E2" AND RACE = "3" THEN GROUP = "1";
else if ETHNICITY = "E2" AND RACE = "5" THEN GROUP = "2";
else IF ETHNICITY = "E1" THEN GROUP = "3";
RUN;

Re: how do you combine two variables into one variable in the same dataset

Ksharp — Fri, 21 Feb 2025 01:44:07 GMT

You could create a new variable by combining these two variable ,and create a format to get this GROUP variable, like:

data have;
set have;
new=catx(' ',ETHNICITY ,RACE );
run;
proc format;
value $ fmt
'E2 3'='1'
'E2 5'='2'
'E1' ='3'
;
run;
data want;
set have;
group=put(new,$fmt.);
run;

Re: how do you combine two variables into one variable in the same dataset

Tom — Fri, 21 Feb 2025 01:52:49 GMT

You need to use a valid data step. Can you fix that?

Your code is only assigning values to the new variable for 3 of the possible 32 combinations of your two variables.

RACE has 7 valid values and I assume could also have missing values, so 8 possible values.

ETHNICITY has 3 valid values and again I assume could also have missing value, so 4 possible values.

8*4= 32 possible combinations.

Do you just want group to be missing for the other 31 possible combinations?

Re: how do you combine two variables into one variable in the same dataset

whyamihere — Fri, 21 Feb 2025 02:13:49 GMT

Thanks for responding. What do you mean by using a valid data step? I have bitten off more than I can chew in this class but really want to finish, so please dumb it down for me. My instructions only want me to analyze these 3 values in relationship to mean age.

Re: how do you combine two variables into one variable in the same dataset

Tom — Fri, 21 Feb 2025 03:25:38 GMT

@whyamihere wrote:

Thanks for responding. What do you mean by using a valid data step? I have bitten off more than I can chew in this class but really want to finish, so please dumb it down for me. My instructions only want me to analyze these 3 values in relationship to mean age.

Data step starts with a DATA statement that names the dataset(s) it is creating.

It needs to have a source of data. If that is an existing dataset then you need a SET statement.

Example:

data want;
  set have;