b-tree

refers to a 'balanced tree' access method which Pick supports throughout the system and in every process which touches the data base.

A 'b-tree' is essentially a system-maintained ordered-list based on any attribute or combination of attributes from items in a file. It allows quick retrieval of items based on this secondary key.

With minor modifications, current Pick applications can make use of 'b-trees'. Any Pick process which modifies an item automatically maintains the tree implicitly. This includes 'update', Pick/BASIC, 'edit', 'copy', 't-load', and 'restore'.

Access retrieves keys from a b-tree when an associated attribute-defining item is used in a 'by' clause, such as 'sort entity by last.name'. 'last.name' points to an attribute which is also referenced by an 'i' (index) processing code in the adi. If more than one 'by' clause is used in the same request, Access ignores the tree and scans the entire file. A 'with' clause with no 'by' clause also uses the tree.

A more flexible approach to incorporating b-trees to data retrieval is to use the Update processor or the Pick/BASIC intrinsic functions like 'key' and 'root'.