Hot Topics
- Updates for SAS 9.2 (March 2008)
See the compatibility calculator to determine whether you need to migrate your libraries to SAS 9.2. Use the PROC MIGRATE calculator to find the right example code for your migration.
- Get a Smarter COMPARE
Validation tools for the MIGRATE procedure now work with CEDA and provide additional details about catalogs, indexes, and integrity constraints.
- Start Planning
See Getting Started for the most efficient pathway through all of the SASŪ9 migration documentation.
- Coexistence
Many SAS®9 products can install side-by-side with SAS 8.2 with very good compatibility. See Peaceful Coexistence: SAS®9 and SAS 8.2 for details.
- Sample Validation SOP
See an example SOP: SAS Installation Qualification / Operational Qualification (PDF) (updated 19 May 2008) for the pharmaceutical industry.
Related Links
Validation
|
Migration
Running Source and Target in Parallel
If validation is of highest concern, you might dedicate a substantial period of time for testing the target installation. This requires you to run the source installation and the target installation in parallel during validation.
See the installation setup step in the Execution phase, where you are advised to set up the target SAS®9 installation in an isolated environment from the source installation. You maintain data and programs in both the source and target environments while you validate that the target installation is running correctly and that the data migration is successful. This is a best practice for many reasons, and it's the most reliable way to provide validation. Customers have reported as many as four or five separate environments:
- source
- backup
- development
- test and QA
- production
Consider the following issues:
- Running source and target in parallel requires additional disk space and CPU time.
- If you plan to have the source and target installations interact, then you must research peaceful coexistence issues.
- If a data set is live (that is, continually updated), then you must either update the two data sets concurrently or designate one data set to be frozen and later discarded.