Reading a mixture of record types: conditional input
Situation: You have a set of raw data that are not laid out consistently. That is, some records are structured one way and others are structured differently. As long as there is something in the records that tells you which is which (and there almost always is), you can read this with SAS. Parenthetically, this situation-records with varying layouts-is difficult or impossible for other data management software to handle.
Before going through a specific example, it will help to understand more of what is going on in the Data step.