SQL or SAS or Neither?

The technical and logical challenges of data preparation and data transformation can differ greatly from project to project. Also, data transformation is quite a separate challenge from database management.

A pragmatic approach requires more than one data manipulation programming language.

If the logical complexity of the required data transformations is not too high, a GUI-oriented application may be best. However, for complex data transformations, a graphical user interface can be inflexible and can get in the way.

SQL works quite well with some data transformations, and can be cumbersome for others. The SAS datastep is a better choice for many types of transformations, but for a many-to-many join, for example, SQL is a better fit. A software development language choice ( Perl, C ) usually requires more development time and maintenance than a specialized tool. Received on Tue Nov 20 2007 - 12:41:58 CST