SELECTIVE INDEX SCAN
Original Publication Date: 1988-Jun-01
Included in the Prior Art Database: 2005-Feb-15
This article describes two new approaches for creating indexes on columns of tables in relational data bases. PRESENT INDEX APPROACH In some relational data base systems, it is possible to create indexes on columns of the tables. The indexes consist of a key value (e.g., employee number, department number etc.; preferably the data most frequently used as an entry in the table) and a pointer (block number) to the corresponding data row, constituting tables. The index table is sorted (ascending or descending) according to the key value. The described indexes exist in, e.g., SQL/DS and DB2 today. However, in order to work efficiently, the index must have a certain characteristic; it must be 'clustered'. This means that the data in the table must be stored in the data base in key sequence.