topic Re: Cleaning a text variable in SAS Programming

Cleaning a text variable

Emma_at_SAS — Tue, 23 Nov 2021 15:44:03 GMT

I have a variable that contains details on how much medicine was administered for each patient during a stay at the hospital and I need to clean some observations.

Examples of clean observations are: 2 GRAMS, 20 OUNCE

for these clean examples, I create two variables one for the quantity and one for the unit.

But some observations are: 4 GRAMS ONLY TODAY, 15 OUNCE ONLY TODAY.

For these examples, I need to remove ONLY TODAY, and then I can separate the quantity and unit values using COMPRESS

quantity=compress(DrugQuantity,"1234567890./","k");
unit=compress(DrugQuantity,"ABCDEFGHIJKLMNOPQRSTUVWXYZ ","k");

How can I remove all those ONLY TODAY parts from my original variable or from my compressed unit variable? I appreciate your suggestions.

Re: Cleaning a text variable

maguiremq — Tue, 23 Nov 2021 15:53:48 GMT

I think this calls for PRXCHANGE. If someone notices an issue with my regex, please feel free to correct it.

Errors may pop up depending on whether the words are consistently spelled the way you showed them.

data have;
var = "4 GRAMS ONLY TODAY, 15 OUNCE ONLY TODAY";
run;

data want;
	set have;
		want_var = prxchange("s/\s*?(ONLY TODAY)\s*?//", -1, var);
run;

Obs var want_var 
1 4 GRAMS ONLY TODAY, 15 OUNCE ONLY TODAY 4 GRAMS, 15 OUNCE

The

\s*?

checks for zero or more spaces before or after the phrase ONLY TODAY.

Re: Cleaning a text variable

Emma_at_SAS — Tue, 23 Nov 2021 16:39:14 GMT

Thanks very much @maguiremq. your solution works perfectly! Thanks