topic Re: Creating one single new variable from multiple variables using IF statement in SAS Procedures

Creating one single new variable from multiple variables using IF statement

Cindarellie — Fri, 01 Nov 2019 13:36:59 GMT

Hello!
I'm trying to create a variable that just pools all of the participants responses into one variable.

For example, there was a question that ask which following fruits do you eat? The participants can check off whether or not they eat each option, where 1=they checked it. So I want to create a variable where all the yes responses are in just one if that make senses.

When I run this code, I don't get any errors and the variable is created, however I noticed through PROC FREQ the frequency doesn't match the original variables-so I am wondering how to make sure it is pulling things correctly or what is happening for it to be getting the wrong number of observations?

I am not sure creating the variable using the if function is possible as that is most familiar method I use to create new variables. I am not familiar with arrays but would would that be better?

if q_apple=1 then
fruits="apple";
else if q_orange=1 then
fruits="orange";
else if q_banana=1 then
fruits="banana";

else if q_grapes=1 then
fruits="grapes";

else if q_watermelon=1 then
fruits="watermelon";

else if q_strawberries=1 then
fruits="strawberries";

else if q_bb=1 then
fruits="blueberries";

else if q_ch=1 then
fruits="cherries";
else if q_other=1 then
fruits="Other";

Re: Creating one single new variable from multiple variables using IF statement

Kurt_Bremser — Fri, 01 Nov 2019 13:18:17 GMT

What if you have more than one fruit checked by a single person?

Re: Creating one single new variable from multiple variables using IF statement

Cindarellie — Fri, 01 Nov 2019 13:31:19 GMT

?I'm sorry I meant to say that the frequency changes

Re: Creating one single new variable from multiple variables using IF statement

Amir — Fri, 01 Nov 2019 13:47:51 GMT

If the output of the proc freq is showing fewer observations then that might be because you have some missing values. The number of missing values might be displayed after all other frequencies, depending upon how you have chosen to write the proc freq code.

It would be helpful if you provided a complete sample data step with sample data that can be created like this and the proc freq code you have used.

Amir.

Re: Creating one single new variable from multiple variables using IF statement

VDD — Fri, 01 Nov 2019 14:26:29 GMT

you might want to look at catx function https://support.sas.com/documentation/cdl/en/lrdict/64316/HTML/default/viewer.htm#a002257076.htm

and create fruits

Re: Creating one single new variable from multiple variables using IF statement

Astounding — Fri, 01 Nov 2019 14:52:49 GMT

If you want PROC FREQ to be able to count everything, your data set needs to contain everything. Look at this code:

if q_apple=1 then
fruits="apple";
else if q_orange=1 then
fruits="orange";

It can only assign one value to FRUITS. If more than one item has been checked, the program won't pick it up. It ends up with APPLE or ORANGE, but not both. Here's the sort of change you need to make before running PROC FREQ:

if q_apple=1 then do;
   fruits="apple";
   output;
end;
if q_orange=1 then do;
   fruits="orange";  
   output;
end;

That way, every selection becomes part of the data that feeds into PROC FREQ. Do NOT use the word ELSE.

Re: Creating one single new variable from multiple variables using IF statement

VDD — Fri, 01 Nov 2019 15:20:04 GMT

data want;
length fruits $256;
set have;
 
if q_apple=1 then
fruits=catx(" ", fruits, "apple");
if q_orange=1 then
fruits=catx(" ", fruits, "orange");
if q_banana=1 then
fruits=catx(" ", fruits, "banana");
if q_grapes=1 then
fruits=catx(" ", fruits, "grapes");
if q_watermelon=1 then
fruits=catx(" ", fruits, "watermelon");
if q_strawberries=1 then
fruits=catx(" ", fruits, "strawberries");
if q_bb=1 then
fruits=catx(" ", fruits, "blueberries");
if q_ch=1 then
fruits=catx(" ", fruits, "cherries");
if q_other=1 then
fruits=catx(" ", fruits, "Other");
run;
proc freq data=want;
tables fruits;
quit;

Re: Creating one single new variable from multiple variables using IF statement

ballardw — Fri, 01 Nov 2019 15:11:54 GMT

@Cindarellie wrote:

Hello!
I'm trying to create a variable that just pools all of the participants responses into one variable.

For example, there was a question that ask which following fruits do you eat? The participants can check off whether or not they eat each option, where 1=they checked it. So I want to create a variable where all the yes responses are in just one if that make senses.

When I run this code, I don't get any errors and the variable is created, however I noticed through PROC FREQ the frequency doesn't match the original variables-so I am wondering how to make sure it is pulling things correctly or what is happening for it to be getting the wrong number of observations?

I am not sure creating the variable using the if function is possible as that is most familiar method I use to create new variables. I am not familiar with arrays but would would that be better?

if q_apple=1 then
fruits="apple";
else if q_orange=1 then
fruits="orange";
else if q_banana=1 then
fruits="banana";

else if q_grapes=1 then
fruits="grapes";

else if q_watermelon=1 then
fruits="watermelon";

else if q_strawberries=1 then
fruits="strawberries";

else if q_bb=1 then
fruits="blueberries";

else if q_ch=1 then
fruits="cherries";
else if q_other=1 then
fruits="Other";

Show example starting data and what the result should look like. Otherwise there is lots of guessing as to what you want.