| 摘要 |
The invention relates to a method for providing statistical data from a data warehouse (410) involving one or more data storage means and one or more processor coupled to the data storage means characterized in that it comprises the steps of:
€¢ defining a plurality of index fields (114) each index field accepting a plurality of index field values;
€¢ creating a plurality of indexed files (432) and hierarchically indexing the files as trees of indexes (300) which includes, for each tree, the steps of:
- hierarchically ordering the index fields (201);
- defining bins, each bin being associated to an index field and gathering one or more index field values for that index field;
- creating indexes by concatenating one or more bins in compliance with the hierarchy of hierarchically ordered index fields to form a sequence of bins, an index comprising only one bin per index field;
- hierarchically indexing the files as trees of indexes (300), each index having zero or more children indexes or and having at most one parent index, so that each child index comprises the same sequence of bins as the index of its parent index plus at least an additional bin associated to an additional index field;
€¢ providing each index with a data container (325, 335) configured to store statistical data, so that each data container is indexed and is directly addressable from within the files hierarchically indexed;
€¢ receiving (436) one or more input files (434) comprised of raw data and updating the data containers with the input files (434), which includes for each input file using one or more processor arranged for performing the steps of:
- identifying and extracting from the raw data at least one attribute to be analyzed through statistics and one or more input file parameters characterizing the attribute;
- creating at least an individual record (620) from the input file (434), each individual record (620) comprising at least an attribute and the one or more input file parameters characterizing said attribute;
- associating each input file parameter with at least an index field;
- establishing a correspondence between each input file parameter and a bin of the at least one index field associated to that input file parameter;
- identifying data containers indexed with the one or more bins that all correspond to an input file parameter of said individual record (620);
- incrementally updating the identified data containers (325, 335) with the at least one attribute of said individual records (620) to obtain statistical data describing the attribute. |
