ADABAS provides a series of system tables which can be used to request data definitions, statistics, and database states. These tables are described in the Reference manual.
In particular, requests can be issued onto these system tables to determine the columns, especially the key columns, of tables.
The columns to be selected should always be specified explicitly in a SELECT statement. The statement in the format SELECT * should be avoided for two reasons:
1. In this way, modifications of the table definition can enter unnoticed into the SELECT statement, which may lead to errors if SQLWARN3 is not checked.
2. In most cases, the values of all columns in a table are not needed in the application. The set of selected columns should be restricted to the necessary size in order to save time and space.
For set requests without an order specification (ORDER BY), the sequence in which the table rows are fetched into the program is not determined. Thus the sequence can differ according depending on whether a secondary index is used or not for access support. The sequence can also change from one ADABAS version to the other. The logic of the database application must therefore not rely on the present, but arbitrary output sequence.
If the key value of the row to be accessed is known, a single row statement should be used instead of multiple processing. It should be noted, however, that these single row statements are extensions to the SQL standard.