Date:         Tue, 28 Jan 2003 22:02:37 -0800
Reply-To:     Floyd Nevseta <f.nevseta@CYBERDUDE.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Floyd Nevseta <f.nevseta@CYBERDUDE.COM>
Organization: http://groups.google.com/
Subject:      Re: How to read multiple values?
Content-Type: text/plain; charset=ISO-8859-1

saqi2000@TOTALISE.CO.UK (Saqi) wrote in message news:<3E379A25@mail.totalise.co.uk>...
> Hi sas gurus,
> I have got a macro called ws_ids and I have 200 data sets and I want this
> macro to extract given employees numbers and put into new data sets. What I
> want to do supply 50 employees id at a time by using following command
>
> %get_ids(123456789)
>
> 50 times. But this overwrites the existing data. And I don't know any otherway
> to do it. Can anyone help please.
>
> Thanks in advance
>
> Tahir

Saqi,

Let me see if I understand your problem correctly. You have a macro,
get_ids, that has only one parameter, a single employee id. The macro,
I assume, queries some data set for this employee id and creates a new
data set from the results of the query. However, you need to query the
data set for 50 employee ids, not just one. Additionally, you are not
just querying one dataset but 200 data sets. I think that summarizes
it.

First of all, forget about using the macro, get_ids. Think about this
for a moment. You must run the same query 50 times per data set. With
200 data sets, the query will execute 10,000 times. That's insane,
regardless of the size of the data sets.

A better approach is to put the 50 employee ids in a data set. Then
you can use SQL with an inner join or a data step with match-merging
to get the resulting observations. Instead of having to execute the
query 50 times, this method requires only one read of the data set. A
simple query to capture this process is shown below.

proc sql;
create table out_ds as
select *
from emp_ids emp, in_ds ds
where emp.emp_id = ds.emp_id;
quit;

You can easily write a macro to encapsulate this query, so you can use
it for each of the 200 data sets.

Hopefully this gives you clearer view of your problem and helps you
develop a better solution.

Regards,
Floyd