X-HYBRIDJOIN for near-real-time data warehousing
BNCOD'11 Proceedings of the 28th British national conference on Advances in databases
Bibliography on data base structures
ACM SIGMIS Database
| Hi-index | 0.00 |
Formulas are derived for the average number of record references required to retrieve a record from a file (a) in case the records are loaded without regard for relative frequency of reference and (b) in case each set of records with a common home address is arranged in order of decreasing frequency of reference. The formulas are first derived under the assumption that the mapping from keys to addresses is "random." Finally, an informal argument is given which suggests the formulas will also hold under a familiar "pseudo-random" mapping based on the use of division, provided the keys have a certain property commonly encountered in practice.