topic Re: Pattern Recognition in SAS Programming

Pattern Recognition

SASSLICK001 — Sat, 11 May 2019 21:45:36 GMT

Hi All

I have a challenge to identify a particular pattern in the data, for example

AB, YZ

AB, AB, YZ

AB, AB, AB, YZ

...

if string AB comes before string YZ, i need to identify these records, but AB can be any number of times (i.e. 1 to n)

Any thoughts?

PeterClemmensen — Sat, 11 May 2019 22:29:07 GMT

So in this example, all three records would be identified, correct?

Shmuel — Sun, 12 May 2019 00:20:01 GMT

Try to check by:

if scan(<variable namr>, -1, ' ,') = 'BZ' and /* last string */

scan(<variable namr>, -2, ' ,') = 'AB'

then ...; else ...;

PGStats — Sun, 12 May 2019 02:49:40 GMT

Regular expressions are all about text patterns.

if prxMatch("/AB[, ]+YZ/i", string) then <do something>;

would match your example pattern. The pattern says: AB followed by one or more comma or a space characters, followed by YZ.

Tom — Sun, 12 May 2019 04:18:09 GMT

If it is really as simple as your example then no need for regular expressions.

index(string,'AB, YZ')