Materialized View Maintenance Methods And Performance Evaluation

Abstract

A data warehouse stores views derived from information which might not reside at the factory. These views are actually called as materialized views. Using these materialized views, user queries may be answered quickly because querying the outside sources of energy where the starting information resides is actually stayed away from. Nevertheless, when the sources shift, the views in the factory may become inconsistent with the starting information and should be looked after. These views possibly could be recomputed from scratch, or maybe incremental maintenance methods could be used to compute the changes to the views as a result of the source changes. A variety of approaches have been recommended for keeping these views. Based on whether the present materialized views in a data warehouse are actually used in the computation of the latest views of theirs in reaction to changes which happen on the remote information source, the existing information warehouse view maintenance methods may be classified into 2 large categories. In a single method, the necessary view updates are actually computed without restricting which base associations might be used. This method reduces up to sixty three % view maintenance time, since maintenance does periodically not for every update. We studied and investigated the materialize view maintenance techniques in data warehousing system. We've suggested an intermediate approach to warehouse maintenance depending on what we call materialized view self maintenance, exactly where the views are incrementally maintained without wearing all the starting relations. Just the actual change is actually computed and stored at the information warehouse site to keep the views.