Method of Identifying Unkeyed Records in a Two-Level Index
Original Publication Date: 1985-Mar-01
Included in the Prior Art Database: 2005-Feb-18
A method is described which allows keyed and unkeyed records to be interspersed in the same dataset even though only one index is used. The relationship of the keyed and unkeyed records is determined by the application program and not the storage access method. In this method the storage access method program uses one two-level index to locate both keyed and unkeyed records. Each record locator contains a key which must be numeric. In this instance the keys are three-byte packed hexadecimal so the highest available key is 999999. Any value above this is invalid except FFFFFF which is a special dummy record and AAAAAA which is the "key" assigned to the key fields of unkeyed records. When the indices are scanned, the key being searched is unpacked and then becomes the search argument.