Query processing for selection and projection using inverted partitioned indexes

Very large data volumes and reading data from different information sources and updating data in local information sources make traditional indexing techniques such as the Record Identification (RID) index unsuitable for data warehouses. Bitmap indexes are a promising technique to replace RID at the selection level of query processing. We introduce a variant of the inverted partitioned index that materializes the secondary key of the base relation to the store primary key instead of RID. Query Processing for selection and projection using the inverted partitioned index can process relevant data at the index level without going through the base relation. We use a disk page for the criteria to analyse the inverted partitioned index compared with related techniques such as full scan, RID index and Bitmap Index. Our new inverted partitioned index is much better than these related techniques