topic SAS perl question - lookbehind in SAS Procedures

SAS perl question - lookbehind

allaboutsas — Tue, 30 Dec 2014 20:33:49 GMT

Is there a way to use lookbehind for a few possible letters as a whole? In the example below, I need ABC, DEF, GHI be found as a whole (ABCDE is ok), but seems like it was taken as each letter. I tried (ABC|DEF|GHI), it’s not working, one word is working. Any suggestions?

data a;

var='ABC 1.89 DEF A YELLOW LINES'; output;

var='ABC 1.89 YELLOW JKLM LINE'; output;

var='9 GHI YELLOW LINE'; output;

var='9 H BLK LINE'; output;

var='9 BC JT GREEN LINE'; output;

run;

data b;

set a;

retain re1;

if _N_=1 then do;

re1 = prxparse('/(?<=[ABC|DEF|GHI] )(YELLOW|BLK|GREEN)(?=\s*LI)/');

end;

if prxmatch(re1,var) then do;

var2 = prxposn(re1,0,var);

end;

run;

proc print data=b; run;

1	ABC 1.89 DEF A YELLOW LINES	1	YELLOW
2	ABC 1.89 YELLOW JKLM LINE	1
3	9 GHI YELLOW LINE	1	YELLOW
4	9 H BLK LINE	1	BLK
5	9 BC JT GREEN LINE	1

Re: SAS perl question - lookbehind

ballardw — Tue, 30 Dec 2014 20:58:37 GMT

Please provide details on what you mean by "not working"? What was your expected output?

Re: SAS perl question - lookbehind

gergely_batho — Tue, 30 Dec 2014 21:12:38 GMT

I think you need () instead of []

prxparse('/(?<=(ABC|DEF|GHI) )(YELLOW|BLK|GREEN)(?=\s*LI)/')

Re: SAS perl question - lookbehind

allaboutsas — Tue, 30 Dec 2014 21:14:33 GMT

For example if I have “re1 = prxparse('/(?<=ABC).*(\bYELLOW|BLK|GREEN\b)(?=\s*LI)/');” in code, so ‘YELLOW’ in “ABC 1.89 DEF A YELLOW LINES” will be captured, but I need to list more possible words, like GHI, so YELLOW in “9 GHI YELLOW LINE”’ can be captured too.

Re: SAS perl question - lookbehind

allaboutsas — Tue, 30 Dec 2014 21:19:41 GMT

gergely! you are right. I thought I tried that, got error earlier, I must have added an extra closed parenthesis. Thank you!

Re: SAS perl question - lookbehind

allaboutsas — Tue, 30 Dec 2014 22:27:42 GMT

Hi, is there a way to look around, either ahead or behind, for example, as long as ABC or GHI is in the string, before or after (\bYELLOW|BLK|GREEN\b) return the color?