Date:         Sun, 12 Nov 2006 10:00:05 -0500
Reply-To:     "Richard A. DeVenezia" <rdevenezia@WILDBLUE.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Richard A. DeVenezia" <rdevenezia@WILDBLUE.NET>
Subject:      Re: Quick and hopefully easy question
Comments: To: sas-l@uga.edu

alves wrote:
> Hi guys.
>
> Need to make a stange merge, and having some problems how to do it (at
> least with datasteps).
>
> I have 2 datasets (treatment and no-treatment) and I need to merge the
> datasets in a way that each observation in the treatment dataset will
> be merged with all no-treatment patients with the same age. some
> treatments patients have the same age and when the age repeats I need
> again all no-treatment patients.
>
> Datasets are pretty big so I trying to find a way that optimises
> speed.

Have you reached some bottleneck or case where a current method takes "too
long".

How big is big?
- How many rows and columns in each table?
- Does one table consistently have more rows than the other ?

Are the tables indexed?

What have you tried, if anything?

Was it anything like a left join?
----------------------------------
data one(index=(focus1=(id1 id2) age));
input id1 $ id2 $ age;
cards;
a xx  25
a yy  30
c tt  25
d tt  35
e zz  40
run;

data two (index=(id age));
input id age;
cards;
1  25
2  25
3  30
4  35
5  35
6  25
7  25
8  30
run;

options msglevel=I;

proc sql;
  create table three as
  select one.*, two.id
  from
    one
  left join
    two
  on
    one.age = two.age
  order by
    id1, id2, id
  ;
quit;
----------------------------------

There are many other ways to join data.

Richard A. DeVenezia
http://www.devenezia.com/