A join can be performed across 16 tables at the most. Thereby the tables are joined step by step, i.e., a join is formed from two tables. Each additional table is then combined with the join result to form another join result.
The procedure according to which a result table is generated can be described in pseudo code such as follows:
Time or space can only be saved when the temporary result tables are as small as possible and when the rows of the tables to be combined to form a new join can be accessed directly.
The ADABAS optimizer therefore tries to put small tables with restrictive conditions at the beginning of the series of tables to be processed in order to obtain small temporary result tables.
The sequence of the tables' specification in the <from clause> of the SELECT statement has no influence on the sequence of processing. The sequence of processing is determined by the ADABAS optimizer.
See also: