SAS/SHARE Cross-Version Issues, SAS 9.2 |
Resolving the Number of Observations and Variables Supported across Versions |
More observations and variables are supported in SAS 9.2 than in SAS 8.
For example:
If a SAS 9.2 client accesses SAS files on a SAS 8 server, the client can access as many observations and variables as the server.
If a SAS 8 client accesses SAS files on a SAS 9.2 server, the client cannot access as many observations and variables as the server can store in the file or supply to the client.
To resolve differences between the number of observations and variables that are supported in a cross-version environment, SAS/SHARE will support the older version. For example, a SAS 9.2 server will not present more variables or observations to a SAS 8 client than the client is capable of accessing. Conversely, SAS 8 clients might not be able to access all the variables or observations in SAS files that are on a SAS/SHARE 9.2 server.
Number of Variables Supported |
The maximum number of variables that can be defined in a SAS data set and accessed by a SAS/SHARE client varies according to the SAS version that is used. The number of variables that can be defined and accessed in a data set is greater in SAS 9.2 than in SAS 8.
Consequently, a SAS/SHARE 9.2 client can access all the variables in a SAS 9.2 data set that is on a SAS 9.2 server. Also, a SAS/SHARE 8 client can access all the variables in a SAS 8 data set that is on a SAS 8 server. However, a SAS/SHARE 8 client can access only a maximum of 32,767 variables in a SAS 9.2 data set that is on a SAS/SHARE 9.2 server. Attempts to exceed this limit will result in failure.
Number of Observations Supported |
The number of observations in a file is a consideration only if a SAS 8 client must access observations by number (for example, if the POINT= option in the SET statement is used). A SAS 8 SAS client that reads through a SAS file sequentially is not limited by the number of observations in the file.
A SAS 9.2 client that accesses SAS files on a SAS 8 server can randomly access any observation.
Copyright © 2007 by SAS Institute Inc., Cary, NC, USA. All rights reserved.