Date:         Thu, 28 Feb 2002 15:53:59 -0500
Reply-To:     Ian Whitlock <WHITLOI1@WESTAT.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Ian Whitlock <WHITLOI1@WESTAT.COM>
Subject:      Re: referencing sas views in sql vs data step
Comments: To: Nick <nick_s_downie@HOTMAIL.COM>
Content-Type: text/plain; charset="iso-8859-1"

Nick,

Don't make problems for your self.  If you don't want, don't get it.  For
your example use

proc sql;
 create view vindex as
      select *
            ,99 as number
        from index ( drop = number ) ;
 quit;

IanWhitlock@westat.com

-----Original Message-----
From: Nick [mailto:nick_s_downie@HOTMAIL.COM]
Sent: Thursday, February 28, 2002 12:39 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: referencing sas views in sql vs data step


Hi, hope u can help - I basically want to create a view of a large
dataset but also to overwrite one of its variables. I know a view
holds as many columns as you define, and that these can be called the
same name, and I know that when this is read into sql it is dealt with
differently than datastep, but how can i create a view that will
ensure a consistent approach, ie value always overwritten regardless
of where the view is used?

thanks

In this case d/set one is altered by two is not. Also NB i need to use
select * as the source dataset contains > 100 variables, and this may
be liable to change.

Test code below..

data index;
input var1 $ number;
cards;
A 2
B 3
C 4
D 9
;
run;


proc sql;
 create view vindex as
      select *
            ,99 as number
        from index;
 quit;

data one;
 set vindex;
run;

proc sql;
 create table two as select * from vindex;
quit;