topic Re: Negative Lookahead? in SAS Programming

Negative Lookahead?

allaboutsas — Wed, 18 Mar 2015 15:29:02 GMT

I’m trying this example below, thought what it does is asserts what immediately follows is not “ dollars”, I expected var2 for the first record “100 dollars” returns null, but instead it returns 10, what should I use if I don’t want any value returned for the first record?

data check;

var='100 dollars'; output;

var='100 pesos'; output;

var='USD100'; output;

var='JPY100'; output;

var='JPYbdl100'; output;

run;

data check2;

set check;

*** create the regular expression only once ***;

retain re1;

if _N_=1 then do;

re1 = prxparse('/\d+(?! dollars)/');

end;

if prxmatch(re1,var) then do;

var2 = prxposn(re1,0,var);

end;

run;

proc print data=check2; run;

Re: Negative Lookahead?

allaboutsas — Wed, 18 Mar 2015 15:32:41 GMT

I can't specify 100 because it could be any number even decimals.

Re: Negative Lookahead?

Reeza — Wed, 18 Mar 2015 15:38:48 GMT

What are you trying to do?

Re: Negative Lookahead?

RW9 — Wed, 18 Mar 2015 15:40:40 GMT

Hi,

Am not sure what you are trying to do. Is it just to get the number part of var? If so then compress with alphas option should work fine:

http://support.sas.com/documentation/cdl/en/lrdict/64316/HTML/default/viewer.htm#a000212246.htm

Re: Negative Lookahead?

allaboutsas — Wed, 18 Mar 2015 16:29:17 GMT

Let me give a better example:

data check;

var="Patient reported 100 c headache 3f and nausea. MD noticed rash."; output;

var="a2 Pt. Rptd. Backache. usd25"; output;

var="b25h of patient reported seeing spots."; output;

var="3g Elevated pulse and 6d labored breathing."; output;

var="Headache."; output;

run;

Yes, trying to get number part of var, but only if it meets the following conditions:

What proceeds is not a or b, example a2 or b25 is not wanted.

What follows is not c or d, example 100 c or 6d is not wanted.

Re: Negative Lookahead?

Haikuo — Wed, 18 Mar 2015 18:31:45 GMT

For you example as is, you need to specify the word boundary for '100', otherwise, '0' is not 'dollars', so '10' meets your criteria and is selected.

data check;

var='100 dollars'; output;

var='100 pesos'; output;

var='USD100'; output;

var='JPY100'; output;

var='JPYbdl100'; output;

run;

data check2;

set check;

*** create the regular expression only once ***;

retain re1;

if _N_=1 then do;

re1 = prxparse('/\d+\b(?! dollars)/');

end;

if prxmatch(re1,var) then do;

var2 = prxposn(re1,0,var);

end;

run;

Re: Negative Lookahead?

allaboutsas — Wed, 18 Mar 2015 18:42:40 GMT

Thanks Hai.kuo. boundary seams working for the ones with spaces, will miss the ones without space. any other suggestions?

Re: Negative Lookahead?

Haikuo — Wed, 18 Mar 2015 18:58:18 GMT

Are you sure about that?

data check;

var='100 dollars'; output;

var='100dollars'; output;

var='100 pesos'; output;

var='USD100'; output;

var='JPY100'; output;

var='JPYbdl100'; output;

run;

data check2;

set check;

*** create the regular expression only once ***;

retain re1;

if _N_=1 then do;

re1 = prxparse('/\d+\b(?! dollars)/');

end;

if prxmatch(re1,var) then do;

var2 = prxposn(re1,0,var);

end;

run;

Re: Negative Lookahead?

allaboutsas — Wed, 18 Mar 2015 19:05:59 GMT

Ok, thanks, you are right. I expanded it a little to include look behind, now how come JPY100 and JPYbd1100 don't return anything? any idea?

data check;

var='100 dollars'; output;

var='100dollars'; output;

var='100 pesos'; output;

var='USD100'; output;

var='USD 100'; output;

var='JPY100'; output;

var='JPYbdl100'; output;

run;

data check2;

set check;

*** create the regular expression only once ***;

retain re1;

if _N_=1 then do;

re1 = prxparse('/(?<!D )\b\d+\b(?! dollars)/');

end;

if prxmatch(re1,var) then do;

var2 = prxposn(re1,0,var);

end;

run;

proc print data=check2; run;

Re: Negative Lookahead?

Haikuo — Wed, 18 Mar 2015 20:41:00 GMT

It is indeed not that easy, The core of your problem is to 'negatively' identify an specific char or string, as it may sound simple , but it actually has two layers while PRX can only handle one at a time. One layer is that you want it to be an Alphabet, another layer is that you don't want it to be certain Alphabet. The following code may not cover all of your ground, but it seems working on your presented data as is:

data check;

var='100 dollars'; output;

var='100dollars'; output;

var='100 pesos'; output;

var='USD100'; output;

var='USD 100'; output;

var='JPY100'; output;

var='JPYbdl100'; output;

run;

data check2;

set check;

if prxmatch('m/usd|dollar/io',var)=0 then

var2=prxchange('s/(^\D*|\D*$)//io',-1,var);

run;

proc print data=check2; run;

Re: Negative Lookahead?

Ksharp — Thu, 19 Mar 2015 12:33:21 GMT

If I understand what you mean.

data check;

var='100 dollars'; output;

var='100dollars'; output;

var='100 pesos'; output;

var='USD100'; output;

var='USD 100'; output;

var='JPY100'; output;

var='JPYbdl100'; output;

run;

data want;

set check;

length w $ 20;

if not prxmatch('/(JPY|JPYbd)\s*\d+|\d+\s*(dollars)/i',var) then w=compress(var,,'kd');

run;

Re: Negative Lookahead?

allaboutsas — Tue, 24 Mar 2015 20:53:25 GMT

Sorry for the late reply. Thank you Hai.kuo and xia keshan. The codes both of you provided work only for this sample, won’t work for the other one I gave, which is closer to my real data:

data check;

var="Patient reported 100 c headache 3f and nausea. MD noticed rash."; output;

var="a2 Pt. Rptd. Backache. usd25"; output;

var="b25h of patient reported seeing spots."; output;

var="3g Elevated pulse and 6d labored breathing."; output;

var="Headache."; output;

run;

As you can see var could have other numbers in it, while I only want those not started with a or b, and not followed with c or d. Is there a way to do it?

Re: Negative Lookahead?

ballardw — Tue, 24 Mar 2015 20:58:14 GMT

What do you want the output to look like for this data sample?

What would you expect for a data line like:

Pulse 110 BP 70 over 30

Re: Negative Lookahead?

allaboutsas — Tue, 24 Mar 2015 21:08:47 GMT

I should return this:

var var2

Patient reported 100 c headache 3f and nausea. MD noticed rash.	3
a2 Pt. Rptd. Backache. usd25	25
b25h of patient reported seeing spots.
3g Elevated pulse and 6d labored breathing.	3
Headache.

Thanks.

Re: Negative Lookahead?

allaboutsas — Tue, 24 Mar 2015 21:11:58 GMT

And what's before and after the number is random, there's not way to specify them.

Re: Negative Lookahead?

Patrick — Tue, 24 Mar 2015 21:43:08 GMT

Is below what you're after?

data check;

var="Patient reported 100 c headache 3f and nausea. MD noticed rash.";

output;

var="a2 Pt. Rptd. Backache. usd25";

output;

var="b25h of patient reported seeing spots.";

output;

var="3g Elevated pulse and 6d labored breathing.";

output;

var="Headache.";

output;

run;

data want;

if _n_=1 then _rx=prxparse('/(^|[^ab\d])(\d+)([^cd\d]|$)/oi');

retain _rx;

set check;

length string $20;

_check=var;

_check=compress(_check);

if prxmatch(_rx,_check) then

do;

string=prxposn(_rx,2,_check);

end;

run;

Re: Negative Lookahead?

allaboutsas — Tue, 24 Mar 2015 22:03:53 GMT

Thank you, Patrick!

Sorry, just one more scenario, for cases like this one below:

var="abc h3 4k Headache.";

There are 2 valid numbers, I'm ok with if 3 is returned, but not 34, any way to fix?

Thanks!

Re: Negative Lookahead?

Patrick — Wed, 25 Mar 2015 07:50:01 GMT

Not really pretty but best I could come up with:

data want;

if _n_=1 then _rx=prxparse('/(^|[^ab\d])(\d+)([^cd\d]|$)/oi');

retain _rx;

set check;

length string $20;

_check=var;

_check=prxchange('s/(\d) +(\d)/\1\|\2)/o',-1,_check);

_check=compress(_check);

if prxmatch(_rx,_check) then

do;

string=prxposn(_rx,2,_check);

end;

run;

Re: Negative Lookahead?

Ksharp — Wed, 25 Mar 2015 12:47:22 GMT

data check;

var="Patient reported 100 c headache 3f and nausea. MD noticed rash."; output;

var="a2 Pt. Rptd. Backache. usd25"; output;

var="b25h of patient reported seeing spots."; output;

var="3g Elevated pulse and 6d labored breathing."; output;

var="Headache."; output;

var="abc h3 4k Headache.";output;

run;

data want;

set check;

length v $ 20;

retain pid;

if _n_ eq 1 then pid=prxparse('/\b[c-z]+\d+\b|\b\d+[abe-z]+\b/i');

call prxsubstr(pid, var, position, length);

if position ne 0 then v = compress(substr(var, position, length),,'kd');

drop pid position length;

run;

Xia Keshan

Re: Negative Lookahead?

allaboutsas — Wed, 25 Mar 2015 15:37:24 GMT

Hi Patrick,

Thank you so much! It’s working mostly, but still something is missing. I have to change my sample to suit better of my real data:

The number wanted has to be a decimal: (\d*\.\d+)
Words don’t want to be in front of the number are ‘abc’, ‘efg’
Letters don’t want to be behind the number are ‘c’, ‘d’

data check;

var="Patient reported .1 c headache 1.3f and nausea. MD noticed rash."; output;

var="abc2.2 Pt. Rptd. Backache. usd2.5"; output;

var="efg2.5h of patient reported seeing spots."; output;

var="1.3g Elevated pulse and 0.6d labored breathing."; output;

var="Headache."; output;

var="abc .3 5 4.5k Headache.";output;

run;

Expected Output:

var string

Patient reported .1 c headache 1.3f and nausea. MD noticed rash.	1.3	Patientreported.1cheadache1.3fandnausea.MDnoticedrash.
abc2.2 Pt. Rptd. Backache. usd2.5	2.5	abc2.2Pt.Rptd.Backache.usd2.5
efg2.5h of patient reported seeing spots.		efg2.5hofpatientreportedseeingspots.
1.3g Elevated pulse and 0.6d labored breathing.	1.3	1.3gElevatedpulseand0.6dlaboredbreathing.
Headache.		Headache.
abc .3 5 4.5k Headache.	4.5	abc.3\\|54.5kHeadache.