topic Re: finding matches for a long list of patterns in SAS Programming

finding matches for a long list of patterns

gzr2mz39 — Fri, 04 May 2018 21:54:59 GMT

I have a long list of patterns that I need to check to see if there are matches in my data.
Which function do you recommend that would be the quickest to create code for checking for a long list of patterns?
Types of patterns I'm interested in matching are:
122Y[digit 0-9]P68A
12811[digit 0-9][digit 0-9]E
13559[digit 0-9][digit 0-9]
65G101[digit 0-9][digit 0-9]
F6364[digit 0-9][digit 0-9]-66[letter A-Z]

Thank you.

Re: finding matches for a long list of patterns

Reeza — Fri, 04 May 2018 22:32:28 GMT

Regex.

https://support.sas.com/rnd/base/datastep/perl_regexp/regexp-tip-sheet.pdf

Re: finding matches for a long list of patterns

gzr2mz39 — Tue, 08 May 2018 20:19:17 GMT

Thank you. This macro seems to work well.

Do you have any more advice with regards to using regex functions with a long list of patterns?

In addition to creating a long list of patterns to check, the list of values I'm checking for these patterns is >700,000.

Any advice for improving efficiency would be appreciated.

%macro prx(pattern);
b=prxparse("&pattern");
if prxmatch(b,serial_number)>0 then check=1;
%mend;

%prx(/^2C1522[\d][\d]$/);

Thank you.

Re: finding matches for a long list of patterns

Reeza — Wed, 09 May 2018 19:29:18 GMT

I'm not that familiar with regex so I would suggest reposting this specifically and asking for efficiency, if you already have not.

There is a way to speed it up so that SAS is not resolving the prx every time but I don't know enough to illustrate it.