topic Re: More efficient way to search for strings in a character field? in SAS Programming

More efficient way to search for strings in a character field?

CaesarS — Thu, 12 Mar 2009 13:50:55 GMT

I'm currently using the code below to search field ABC for one of several text strings; if any of them appear, I set NEW_FIELD to a common result. Is there a better way to do this? In plain english, I'm looking for a statement that says, if ABC contains any of the following: 'STRING1', . . . 'STRING6' then NEW_FIELD = 'RESULT1'

IF INDEX(ABC, 'STRING1') > 0 THEN NEW_FIELD = 'RESULT1';
ELSE IF INDEX(ABC, 'STRING2') > 0 THEN NEW_FIELD = 'RESULT1';
ELSE IF INDEX(ABC, 'STRING3') > 0 THEN NEW_FIELD = 'RESULT1';
ELSE IF INDEX(ABC, 'STRING4') > 0 THEN NEW_FIELD = 'RESULT1';
ELSE IF INDEX(ABC, 'STRING5') > 0 THEN NEW_FIELD = 'RESULT1';
ELSE IF INDEX(ABC, 'STRING6') > 0 THEN NEW_FIELD = 'RESULT1';

Re: More efficient way to search for strings in a character field?

sbb — Thu, 12 Mar 2009 14:34:49 GMT

The INDEX function appears to be the most efficient, as compared to the recently added FIND function.

Also, using your coding example, I also tested using OR condition rather than IF THEN ELSE logic and again INDEX was more efficient, say with a 1 million DO loop iteration with an OUTPUT.

Scott Barry
SBBWorks, Inc.

Re: More efficient way to search for strings in a character field?

deleted_user — Thu, 12 Mar 2009 17:52:34 GMT

Maybe indexc is more efficient in this context.

if indexc(abc,'STRING1',...,'STRING6') > 0 then NEW_FIELD = 'RESULT1';

Re: More efficient way to search for strings in a character field?

sbb — Thu, 12 Mar 2009 20:14:56 GMT

The function INDEXC applies to a character (one or more specified as argument #2) search within a string (argument #1), not a string of characters to search for.

Scott Barry
SBBWorks, Inc.