Querying Hierarchical Data in Very Large Databases

Hierarchical data, such as Partially Ordered Set (POSET) is tremendously used in relational databases, especially in data mining and data warehouse based-applications. Unfortunately, SQL (Structured Query Language) does not effectively support hierarchical data structure to manage this sort of data, for example, in Oracle, a CONNECT BY operator is used to query data organized into trees, however; such operator is not efficient in the presence of very large databases with respect to the response time of SQL queries. Furthermore, SQL lacks of functionalities to manipulate data organized into POSET such as the POSET operators (UB, LB, JOIN, MEET). Therefore, the main objective of our works is to propose a new mechanism to manipulate POSET data using efficient implementation techniques, high performance hardware to enable POSET operators to be able to handle on very large databases.