Methods and apparatus for aggregating sparse data

A method for creating a composite-join hierarchy for use in aggregating a multi-dimensional array of data values, said multi-dimensional array of data values being organized in at least one dimension in accordance with a plurality of hierarchical relationships, said method comprising, in combination, the steps of:

• merging said plurality of hierarchical relationships into a single hierarchical relationship,
• storing an executable program for processing the data in an input worklist representing a data array of n dimensions which is organized into hierarchical levels in at least one of those dimensions, said program producing a sublist for every combination of hierarchical levels represented in said data array,
• storing an initial worklist containing data representing said multidimensional array of data values,
• employing an iterative routine for first executing said program to first process said initial worklist to create result sublists, and to thereafter repeatedly execute said program to complete said composite join hierarchy by processing each result sublist produced by said program until no further unprocessed result sublist that contains data remains.

A method for creating a composite-join hierarchy for use in aggregating data in a multi-dimensional array, said multi-dimensional array of data values being organized in at least one dimension in accordance with a plurality of hierarchical relationships, said method comprising, in combination, the steps of:

• merging said plurality of hierarchical relationships into a single hierarchical relationship,
• storing an executable program for subdividing an input data worklist representing a said data array of n dimensions into sublists each representing a data array of (n−1) dimensions forming part of said composite-join hierarchy,
• storing an initial worklist containing data representing said data in a multi-dimensional array,
• employing an iterative routine for first executing said program to first process said initial worklist into to create result sublists, and to thereafter repeatedly execute said program to process each result sublist produced by said program until no further unprocessed result sublist containing data remains to complete the creation of said composite-join hierarchy.

A method for creating a composite-join hierarchy for use in aggregating a multi-dimensional array of data values, said multi-dimensional array of data values being organized in at least one dimension in accordance with a plurality of hierarchical relationships, said method comprising, in combination, the steps of:

• merging said plurality of hierarchical relationships into a single hierarchical relationship,
• storing an executable program for processing the data in an input worklist representing a data array of n dimensions which is organized into hierarchical levels in at least one of those dimensions, said program producing a sublist for every combination of hierarchical levels represented in said data array,
• storing an initial worklist containing data representing said multidimensional array of data values,
• employing an iterative routine for first executing said program to first process said initial worklist to create result sublists, and to thereafter repeatedly execute said program to complete said composite join hierarchy by processing each result sublist produced by said program until no further unprocessed result sublist that contains data remains.