Closest date matching without duplicates

[Sujithpeta]

I've two data sets with patient drug claim date and administration date, I want to map the drug to administration date without duplication of administration date (i.e. one administration date should only be for one drug date) and the hierarchy is to match the admin date happening on the same day as drug date as it's more reliable. If there is no same days admin date then find the closest date but making sure already matched admin dates aren't used again.

Have:

Patient   DrugDate 

A          01/05/2017     

A          01/18/2017

A          01/31/2017   

A          02/28/2017  

Patient   AdminDate

A          01/31/2017   

A          02/28/2017  

Want:

Patient   DrugDate      AdminDate

A          01/05/2017     

A          01/18/2017

A          01/31/2017    01/31/2017

A          02/28/2017    03/17/2017

Appreciate your help. Thanks 

[mkeintz]

What is your rule if you have 2 consecutive drug dates that are unmatched, and you have an admin date exactly half-way between them?

[Ksharp]

As Mike said there are too many scenario you need consider about .

data drug;
input Patient $  DrugDate : mmddyy10.;
format DrugDate mmddyy10.;
cards;
A          01/05/2017     
A          01/18/2017
A          01/31/2017   
A          02/28/2017  
;
 
data admin;
input Patient $  AdminDate : mmddyy10.;
format AdminDate mmddyy10.;
cards;
A          01/31/2017   
A          03/17/2017  
;

data temp;
set drug(in=ina) admin(rename=(AdminDate=DrugDate) in=inb);
by Patient DrugDate;
a=ina;b=inb;
run;
data want;
 merge temp temp(firstobs=2 keep=Patient DrugDate b
 rename=(Patient=_p DrugDate=_d b=_b));
if Patient=_p and a and _b then AdminDate=_d;
if a;
format AdminDate mmddyy10.;
keep Patient DrugDate AdminDate;
run;