High performance in memory data cubes

A computer-implemented method for adding a data point to a multidimensional database having regions, the method comprising:

• transforming, with at least one computer, the data point into a new region;
• for each previously existing region in the multidimensional database:
• determining, with the at least one computer, if the new region differs from the previously existing region by only one dimension;
• merging, with the at least one computer, the previously existing region into the new region if the new region differs from the previously existing region by only one dimension;
• and
• repeating said determining and said merging for each previously existing region in the multidimensional database;
• wherein said merging comprises a computing platform storing data in a database;
• wherein each particular region in the multidimensional database is stored as a separate structure that comprises:
• (a) a value list consisting only of values that (i) were determined to be within the particular region and (ii) are contained in populated non-null cells of the multidimensional database, and
• (b) for each particular dimension in the multidimensional database, a separate coordinate list that only contains, for each particular value in the value list, a coordinate that identifies a position, in the particular dimension, of a particular cell that contains the particular value;
• wherein the step of merging the previously existing region into the new region comprises:
• finding, with the at least one computer, the sole dimension along which the previously existing region differs from the new region;
• for each particular coordinate in a first coordinate list that (i) contains only coordinates from the sole dimension and (ii) is in a first structure that corresponds to the new region, copying, with the at least one computer, the particular coordinate into a second coordinate list that (i) contains only coordinates from the sole dimension and (ii) is in a second structure that corresponds to the previously existing region;
• sorting, with the at least one computer, the second coordinate list after said copying is complete;
• and
• rearranging, with the at least one computer, values in a first value list that is contained in the second structure;
• wherein rearranging the value in the first value list associates values in the first value list one-to-one with points that are represented by a cross-product of at least the second coordinate list and one or more other coordinate lists that correspond to dimensions other than the sole dimension.

A computer-implemented method for merging a previously existing region of a multi-dimensional database into a new region, the method comprising:

• finding, with at least one computer, a sole dimension of the multi-dimensional database along which the previously existing region differs from the new region;
• for each particular coordinate in a first coordinate list that (i) contains only coordinates from the sole dimension and (ii) is in a first structure that corresponds to the new region, copying, with the at least one computer, the particular coordinate into a second coordinate list that (i) contains only coordinates from the sole dimension and (ii) is in a second structure that corresponds to the previously existing region;
• rearranging, with the at least one computer, values in a first value list that is contained in the second structure;
• wherein rearranging the values in the first value list associates values in the first value list one-to-one with points that are represented by a cross-product of at least the second coordinate list and one or more other coordinate lists that correspond to dimensions other than the sole dimension.