摘要 |
<p>A method of searching for sets of data in a very large columnar database is described. The method comprises carrying out a first query by querying the very large columnar database for a first desired set of data, the database having a range of possible data storage locations; creating a compressed columnar representation of the locations within the very large columnar database of members of the first desired set of data; wherein the compressed columnar representation comprises a plurality of representations covering the entire range of the very large database, wherein each representation provides a descriptor describing a distribution of the desired data results within the sub-range of the entire range; storing the compressed columnar representation in a data cache; and applying a logical operation on first desired set of data stored within the very large columnar database by using the compressed columnar representation stored in the data cache in preference to data stored within the very large columnar database.</p> |