⨳ the positive side of keeping all your eggs in one basket ⨳

2012 Jan 05, Thursday

in any work environment with multiple SAS users, the are likely common data sources used by all or most users. how do you go about deciding where to define the connections to these common data sources?

here’s the most common options for placing libname definitions:

1. within each script,
2. within a script for a specific project/data source,
3. within each user’s autoexec file, or
4. within a common autoexec file.

if integrated security is available, or if all users can be identified with &SYSUSERID and share the same password, or if all users share the same credentials, you’d prefer option 4. lower numbered options have the issue of keeping definitions to the same data source across multiple locations. (i’d argue that options 3 & 4 are the only reasonable options.)

changes to data source connections further complicates things, but the impact can be enormous if options 1 or 2 was used. you’ll put your work on hold while you scour old code making modifications to the numerous libname definitions; i’ll make a handful of changes in 2 files and get actual work done.