Relational Data Index Consolidation
Michal Kvet · 2021
A database index is an important object associated with the table to provide an additional layer for data access. By using the index, it is not necessary to scan the table block by block to locate relevant data rows. On the contrary, reflecting the usage of the tree structure it is possible to search for a record with logarithmic complexity. The limitation of the index structure is identified by the covering of the whole data set. If such a requirement cannot be ensured, the index method cannot be used, whereas some data portion does not need to be present in the result set, although all query conditions are passed. A typical example of such a problem is the NULL definition associated with the object or attribute itself. This paper deals with the existing solutions based on various transformation modules and proposes its architecture extending the index by node pointing to the undefined values.