pull date from character field

[shuchidxt_gmail_com]

Hi,

I am trying to extract only distinct date from the below character variable.

Example:

variable abc

has below values and wanted to extract only 3 below distinct dates.

final.case_20190712

final.incident_20190405

final.abcd_20190405

final.trans_tarn_20190406

final.incident

I

[Reeza]

Seems like those are the only numbers so you can consider using COMPRESS with the 'kd' modifiers to keep only digits and then use INPUT() to convert that a numeric SAS date.

[novinosrin · Accepted solution]

If the set of words is alpanumeric, 

This deals with delim

data have;
input abc $30.;
cards;
final.case_20190712
final.incident_20190405
final.abcd_20190405
final.trans_tarn_20190406
final.incident
;
 
data want;
set have;
k=findc(abc,'_','b');
if k>0 then date=substr(abc,k+1);
drop k;
run;

[novinosrin]

@shuchidxt_gmail_com  Oh you want distinct dates

proc sql;
create table want as
select distinct ifc(findc(abc,'_','b')>0,substr(abc,findc(abc,'_','b')+1),' ') as dates
from have
having dates>' ';
quit;

[shuchidxt_gmail_com]

Thanks All. it worked

[Ksharp]

Just for fun.

data have;
input abc $30.;
temp=scan(abc,-1,'_');
if prxmatch('/^\d{8}$/',strip(temp)) then want=temp;
cards;
final.case_20190712
final.incident_20190405
final.abcd_20190405
final.trans_tarn_20190406
final.incident
;