topic Re: Perl expression in SAS Procedures

Perl expression

ybz12003 — Fri, 23 Jun 2017 17:35:08 GMT

Hello:

Could someone explain what are (?<!\d) and d{4}(?!\d) meaning in the following codes? Thanks.

http://analytics.ncsu.edu/sesug/2012/CT-03.pdf

data SSN ;

input SSN $20. ;

datalines ;

123-54-2280

#987-65-4321

S.S. 666-77-8888

246801357

soc # 133-77-2000

ssnum 888_22-7779

919-555-4689

call me 1800123456

;

run ;

proc sql feedback ;

select ssn

, prxchange( ‘s/(?<!\d)\d{3}[-_]?\d{2}[-_]?\d{4}(?!\d)/xxxxxxxxx/io’, -1, ssn)

as ssn2

from ssn

;

quit ;

Re: Perl expression

ballardw — Fri, 23 Jun 2017 17:57:22 GMT

Do you have any of the SAS documentation available?

Re: Perl expression

ybz12003 — Fri, 23 Jun 2017 18:00:58 GMT

From the link I provide below, it states that

The negative look-behind (i.e. (?<!\d) ) and negative look-ahead assertions (i.e. (?!\d) ) are non-capturing.

However, I still don't get what the code means.

Re: Perl expression

ChrisHemedinger — Fri, 23 Jun 2017 18:03:12 GMT

Here's how you can find out.

Visit https://regex101.com.

Paste in the main expression:

(?<!\d)\d{3}[-_]?\d{2}[-_]?\d{4}(?!\d)

See the Explanation field.

Re: Perl expression

ybz12003 — Fri, 23 Jun 2017 18:10:55 GMT

That is awesome tool! Thank you so much, Chris.

Re: Perl expression

kiranv_ — Fri, 23 Jun 2017 18:22:02 GMT

(?<!\d) --negative look behind

(?!\d) --negative lookahead

This values are used to check whether a particular characters are present along with other characters and hence they are called as zero value assertions.

?< symbol indicates look behind, which is to check wether a particular value is there in front of the value you are looking for

! symbol represents not equal to AKA negative

Let us use easier example. if you values goathair, cowhair, cowsomething and you want to find value hair

and also do not want to have goat infront of hair then you something like this (?<!goat)hair

(?<!\d) indicates that you donot want a digit(\d) in the front of your expression given

?! symbol indicates lookahead, which is to check wether a particular value is there in back of the value you are looking for

! symbol represents not equal to AKA negative

Let us take an easier example. if you want to find value of hairyboy goodboy hairykid and you want to find variable with hair

and also do not want to pick hairyboy but the value should have hairy infront then you do something like this hairy(?!boy)

\d{4}(?!\d) --- \d{4} means a digit repeated 4 times after value and there is no digit after that.

I tried my best to explain, please let me know if something is unclear, confusing or wrong

Re: Perl expression

ybz12003 — Fri, 23 Jun 2017 19:12:16 GMT

Thank you very much for your detail explaination, Kiranv_

Re: Perl expression

SAS_inquisitive — Sun, 09 Jul 2017 15:41:49 GMT

@kiranv_ can you explain this?

been*? is matching to "bee". However, been+? is matching to "been".

I know *? matches previous elenment zero or more times. I thought been*? should match to "been".

Thanks!

Re: Perl expression

kiranv_ — Mon, 10 Jul 2017 00:51:01 GMT

? has difference in meaning when it is infront of + and *

+ and * are greedy. It can match condition as long as there is possibility

say the word is "beennnnnnnnnn" , if you say been+ (+ is 1 or more) then it can match "beennnnnnnnnn"

? infront of + and * makes it non greedy.

+? and *? are non greedy. Non greedy means the search stops one it fulfills the least condition. for +? it is one and for *? it is zero)

say the word is "beennnnnnnnnn" and if you instruct regex engine to find been+? then it will pick ups "been" (n is there only once . non greedy here means once it picks first n it stops)

say the word is is "beennnnnnnnnn"if you instruct regex engine to find been*? it picks up "bee"(n can be zero and it stops immediately as non greedy. non greedy here means it can stop without seeing it)

Re: Perl expression

SAS_inquisitive — Mon, 10 Jul 2017 18:17:29 GMT

@kiranv_ this makes sense. Thanks !