1300 variables and 712 observations is what many of us would call a
small data set. A huge data set would have 1300 variables and
Maybe we could start using the same terms for data sets that are used
for olives - the smallest size I've seen sold is "large", and the scale
moves up through extra large, jumbo, extra jumbo, giant, colossal, super
colossal, and mammoth to "super mammoth". (The smaller sizes, by the
way, are superior, brilliant, and fine, with the smallest size being
Assuming that you mean that you want to replace the numeric missing
values, you could use this code in a data step:
array missval _numeric_;
do over missval;
if missing(missval) then
missval = -9;
Untested, but free, code.
Manager, Technical Development
METRICS Department, First Health
West Sacramento, California USA
>>> "Hollie Rich" <hx97403@HOTMAIL.COM> 09/13/2002 12:43 PM >>>
Dear SAS-L users;
I have a huge data set (about 1300 variables and 712 observations). I
to replace the all missing vaules in data set with -9. I wonder is
easy way to do?
>From: Dale McLerran <stringplayer_2@YAHOO.COM>
>Reply-To: Dale McLerran <stringplayer_2@YAHOO.COM>
>Subject: Re: X and Z matrices in PROC MIXED
>Date: Tue, 28 May 2002 09:11:20 -0700
>I don't know of any way to get the design matrices output directly
>from proc mixed. However, you ought to be able to produce the design
>matrices by running PROC GLMMOD. It's sole purpose is to construct
>design matrices. I don't have time to dummy up an example of how
>to use GLMMOD for this purpose. However, let me just point out that
>you will probably want to place random effects and fixed effects
>all on the MODEL statement. GLMMOD mimics the construction of the
>procedure GLM. With PROC GLM, you place all fixed and random effects
>on the MODEL statement. I believe that you will want to do the same
>if you run GLMMOD to generate the design matrices.
>--- knissine@REMOVETHIS.CC.JYU.FI wrote:
> > Hi all
> > Is there any convenient way to extract the
> > X and Z matrices used in a PROC MIXED run
> > into SAS data sets? According to the Online Doc
> > ODS OUTPUT is of no help here.
> > Or, is there any way to create X and Z matrices
> > e.g. in IML and give them to MIXED as input?
> > The point is that I want to carry out some further
> > calculations based on the model estimated by MIXED.
> > Thanks,
> > Kari
> > --
> > Kari Nissinen
> > Department of Mathematics and Statistics
> > University of Jyväskylä, Finland
>Fred Hutchinson Cancer Research Center
>Ph: (206) 667-2926
>Fax: (206) 667-5977
>Do You Yahoo!?
>Yahoo! - Official partner of 2002 FIFA World Cup
Chat with friends online, try MSN Messenger: http://messenger.msn.com