Summary

This chapter provided an overview of planning the ETL approach and creating transformation code. First, we selected native variables to include in our SAS data warehouse, then designed transformed variables to derive during ETL. We used PROC FREQ and PROC UNIVARIATE to study our native variables and make good design decisions about our transformed variables. We documented our decisions in a data dictionary, which we then used as a guide when we created our transformation code. We used SAS data steps to create grouping variables and two-state flags and recoded continuous variables. We checked the variables for accuracy in recoding as we created them, and then we exported a final analytic dataset.

These skills are important to know when running a data warehouse. Making a data dictionary is a great skill to have for planning variables for transformation, as well as for keeping documentation about both native and transformed variables. It is helpful to become adept at using...