topic Re: Wanting to Only Keep Response that could be in Multiple Columns in New SAS User

Wanting to Only Keep Response that could be in Multiple Columns

Pauliet123 — Wed, 07 Feb 2024 14:51:32 GMT

Hello,
I want to only keep observations with a specific value that could be found in multiple columns. Example:
Person 1st 2nd 3rd
A Pizza. Chicken. Fruit
B. Chicken. Pizza. Fruit
C. Fruit. Chicken. Fish

I want to only keep observations where pizza is a listed option regardless of what column it is in. There are 10+ columns so I didn’t want to do an if then for all those columns. Thank you so much!

Re: Wanting to Only Keep Response that could be in Multiple Columns

yabwon — Wed, 07 Feb 2024 14:58:56 GMT

data have;
input Person $1. (_1st _2nd _3rd) (:$10.); /* _ added because variable can't atar with a number*/
cards;
A Pizza. Chicken. Fruit.
B Chicken. Pizza. Fruit.
C Fruit. Chicken. Fish.
;
run;
proc print;
run;

data want;
  set have;
  array food[*] _1st--_3rd;
  if "Pizza." in food;
run;
proc print;
run;

Re: Wanting to Only Keep Response that could be in Multiple Columns

PaigeMiller — Wed, 07 Feb 2024 16:11:03 GMT

Having a period after the value in columns 1 and 2, but not in column 3, seems to complicate this.

@Pauliet123 Is that really representative of the actual data? Or is that something you just put in there while typing your original question and doesn't represent the actual data?

If the actual data has periods in columns 1 and 2 but not in column 3, then the solution from @yabwon fails if "Pizza" (with no period) appears in column 3.

As always, good data layouts with repeatable formatting of the data makes programming easier. Unusual data layouts and/or inconsistent formatting of the data make programming harder.