The following figure shows how SAS data sets are accessed through an engine.
Your data is stored in files for which SAS provides an engine. When you specify a SAS data set name , the engine locates the appropriate file or files.
The engine opens the file and obtains the descriptive information that is required by SAS, for example, which variables are available and what attributes they have, whether the file has special processing characteristics such as indexes or compressed observations, and whether other engines are required for processing. The engine uses this information to organize the data in the standard logical form for SAS processing.
This standard form is called the SAS data file , which consists of the descriptor information and the data values organized into columns (variables) and rows (observations).
SAS procedures and DATA step statements access and process the data only in its logical form. During processing, the engine executes whatever instructions are necessary to open and close physical files and to read and write data in appropriate formats.
Data that is accessed by an engine is organized into the SAS data set model, and in the same way, groups of files that are accessed by an engine are organized in the correct logical form for SAS processing. Once files are accessed as a SAS data library, you can use SAS utility windows and procedures to list their contents and to manage them. See Chapter 26, 'SAS Data Libraries,' on page 467 for more information about SAS data libraries. The following figure shows the relationship of engines to SAS data libraries.