Date: Tue, 16 Dec 2008 21:53:44 -0500
Reply-To: "Howard Schreier <hs AT dc-sug DOT org>"
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: "Howard Schreier <hs AT dc-sug DOT org>"
Subject: Re: Compare and read values into master data table
On Tue, 16 Dec 2008 16:10:10 -0600, Mary <mlhoward@AVALON.NET> wrote:
>The user first needs to let us know how the file is intended to be used; if it
>is to be used for statistics, then it might be worth the effort. I would be
>willing to give it a try, but only if the user lets us know that they are going
>to nearly always use the data in a horizontal structure.
>The point being that we need to know how the user is going to use the
>file in order to best advise how to structure it, not just blanketly tell the
>use that they are wrong to want it in a horizontal fashion, as you seemed to
>be doing. Do you do much statistical analysis?
>Certainly a time series analysis would require the data horizontally.
> ----- Original Message -----
The documentation on "Standard Form of a Time Series Data Set" (see
or http://tinyurl.com/67j9xm) states:
... A time series data set in standard form has the following characteristics:
* The data set contains one variable for each time series.
* The data set contains exactly one observation for each time period.
* The data set contains an ID variable or variables that identify the time
period of each observation ...
Most SAS/ETS procedures that process time series expect the input data set
to contain time series in this standard form, and this is the simplest way
to store time series in SAS data sets ...
You can incorporate cross-sectional dimensions with BY groups, so that each
BY group is like a standard form time series data set. This method is
discussed in the section Cross-Sectional Dimensions and BY Groups.
You can interleave time series, with several observations for each time
period identified by another ID variable. Interleaved time series data sets
are used to store several series in the same SAS variable. Interleaved time
series data sets are often used to store series of actual values, predicted
values, and residuals, or series of forecast values and confidence limits
for the forecasts. This is discussed in the section Interleaved Time Series.
All of this describes long-and-narrow, normalized structure.