Augmented order preserving minimal perfect hash functions for very large digital libraries

• Authors: Amjad M. Daoud;Hussain AbdelJaber;Jafar Ababneh
• Venue: CIT'11 Proceedings of the 5th WSEAS international conference on Communications and information technology
• Year: 2011

Rapid access to information is essential for a wide variety of retrieval systems and applications. Hashing has long been used when the fastest possible direct search is desired, but was considered an exotic [15] and not appropriate when sequential or range searches are also required. To change that, we extend order preserving perfect hash [11] functions to handle sequential access to lexicographically ordered records. To implement this, the lexicographically sorted key set is compressed with prefix compression and stored in blocks or pages in a similar fashion to the leaf pages of a B+ tree with prefix compression [9]. The block number and the key prefix offset within the block are combined to form the key address. The key address is augmented with a signature of the prefix to form the key data. The key data is blended into the function specification to produce an augmented order preserving minimal perfect hash function. Our algorithm uses the bipartite graph approach to avoid degenerate edges problems. It relaxes the acyclic requirement in random graphs presented in [7] and can tolerate the presence of cyclic components. Moreover the algorithm is designed to avoid the conditions described in [7] that make the Fox, Chen, Daoud, and Heath approach [11] exponential. Examples of these conditions are given along with how the algorithm overcomes them. The algorithm produces OPMHFs with much higher success rates than the acyclic hypergraph approach [7] and mostly from the first trial.