System and method for determining driving factor in a data cube

• US 8,234,298 B2
• Filed: 07/25/2007
• Issued: 07/31/2012
• Est. Priority Date: 07/25/2007
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Winter, Stewart James, Westman, Randy Mark, Reid, Murray John, Leikucs, Andrew Alexander, MacCulloch, William Todd
• Primary Examiner(s): Jalil, Neveen Abel
• Assistant Examiner(s): Peach, Polina
• Application Number:
  US11/881,149
• Publication Number:
  US 20090030915A1
• Time in Patent Office: 1,833 Days
• Field of Search: None
• US Class Current: 707/776
• CPC Class Codes: G06F 16/283 Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP

Abstract

Method and system for determining a driving factor for a data value of interest in a multidimensional database, by collecting a context for the data value of interest in the multidimensional database. The data value of interest has dimensional levels with dimensional members outside the drill path of the data value of interest. The dimensional levels are enumerated in a list. A query using the dimensional members of the dimensional level is executed. A variance is calculated for the set of query results. A driving factor for the data value of interest is determined based on the variance. The driving factor is added to the context of the data value of interest.

Claims

1. A method for determining a driving factor for a data value of interest in a multidimensional database, the method comprising:

collecting a context for the data value of interest in the multidimensional database, wherein the data value of interest having a first plurality of dimensions and a drill path, wherein the context comprises an intersection of the first plurality of dimensions, wherein the drill path comprises a first plurality of dimensional levels connected to the first plurality of dimensions, wherein a second plurality of dimensions not directly connected to the first plurality of dimensions comprises dimensions outside the context and outside the drill path, wherein a second plurality of dimensional levels comprise dimensional levels connected to the second plurality of dimensions, the first and the second pluralities of dimensional levels each having dimensional members;

applying the context of the data value of interest to each of the second plurality of dimensional levels to generate a query for each of the second plurality of dimensional levels, wherein, for each of the second plurality of dimensional levels, the query for that dimensional level, when executed, queries the intersection of

1) the first plurality of dimensions, and

2) each of the dimensional members of that dimensional level;

executing, by one or more computing devices, the query for each of the second plurality of dimensional levels using the dimensional members of each of the second plurality of dimensional levels to generate a set of data values for each of the second plurality of dimensional levels, wherein each data value of the set of data values is calculated by at least applying a value of one of the dimensional members of one dimensional level of the second plurality of dimensional levels to the data value of interest, wherein a subset of the second plurality of dimensional levels outside the context and the drill path is used for executing the query;

calculating a variance of the set of data values generated for each dimensional level of the second plurality of dimensional levels;

identifying one of the dimensional levels of the second plurality of dimensional levels as having a greatest variance of the calculated variances;

selecting the identified dimensional level as the driving factor;

generating a new data value of interest having a new context that comprises an intersection of the first plurality of dimensions and a dimension of the driving factor;

selecting the new data value of interest;

removing levels involved with the new data value of interest from the second plurality of dimensional levels;

determining a new driving factor; and

adding the new driving factor to the context.

2. The method of claim 1, further comprising outputting dimensional members for the driving factor in a cross-tabulation.

3. The method of claim 1, wherein calculating a variance determines the minimum and maximum values of the set of data values resulting from the query.

6. The method of claim 1, further comprising enumerating the second plurality of dimensional levels outside the context and outside the drill path in a list.

4. A computer-readable storage device comprising instructions for causing one or more programmable processors to determine a driving factor for a data value of interest in a multidimensional database, wherein the instructions to determine the driving factor further comprise instructions for causing the one or more programmable processors to:

collect a context for the data value of interest in the multidimensional database, wherein the data value of interest having a first plurality of dimensions and a drill path, wherein the context comprises an intersection of the first plurality of dimensions, wherein the drill path comprises a first plurality of dimensional levels connected to the first plurality of dimensions, wherein a second plurality of dimensions not directly connected to the first plurality of dimensions comprises dimensions outside the context and outside the drill path, and wherein a second plurality of dimensional levels comprise dimensional levels connected to the second plurality of dimensions, the first and the second pluralities of dimensional levels each having dimensional members;

apply the context of the data value of interest to each of the second plurality of dimensional levels to generate a query for each of the second plurality of dimensional levels, wherein, for each of the second plurality of dimensional levels, the query for that dimensional level, when executed, queries the intersection of

1) the first plurality of dimensions, and

2) each of the dimensional members of that dimensional level;

execute a query for each of the second plurality of dimensional levels using the dimensional members of each of the second plurality of dimensional levels to generate in a set of data values for each of the second plurality of dimensional levels, wherein each data value of the set of data values is calculated by at least applying a value of one of the dimensional members of one dimensional level of the second plurality of dimensional levels to the data value of interest, wherein a subset of the second plurality of dimensional levels outside the context and the drill path is used for the execution of the query;

calculate a variance of the set of data values generated for each dimensional level of the second plurality of dimensional levels;

identify one of the dimensional levels of the second plurality of dimensional levels as having a greatest variance of the calculated variances;

select the identified dimensional level as the driving factor;

generate a new data value of interest having a new context that comprises an intersection of the first plurality of dimensions and a dimension of the driving factor;

select the new data value of interest;

remove levels involved with the new data value of interest from the second plurality of dimensional levels;

determine a new driving factor; and

add the new driving factor to the context.

7. The computer-readable storage medium of claim 4, further comprising instructions for causing the one or more programmable processors to enumerate the second plurality of dimensional levels outside the context and outside the drill path in a list.

5. A system for determining a driving factor for a data value of interest in a multidimensional database, the system comprising:

a business intelligence analysis server device, wherein the business intelligence analysis server device is configured to;

collect a context for the data value of interest in the multidimensional database, the data value of interest having a first plurality of dimensions and a drill path, wherein the context comprises an intersection of the first plurality of dimensions, wherein the drill path comprises a first plurality of dimensional levels connected to the first plurality of dimensions, wherein a second plurality of dimensions not directly connected to the first plurality of dimensions comprises dimensions outside the context and outside the drill path, and wherein a second plurality of dimensional levels comprise dimensional levels connected to the second plurality of dimensions, the first and the second pluralities of dimensional levels each having dimensional members;

apply the context of the data value of interest to each of the second plurality of dimensional levels to generate a query for each of the second plurality of dimensional levels, wherein, for each of the second plurality of dimensional levels, the query for that dimensional level, when executed, queries the intersection of

1) the first plurality of dimensions, and

2) each of the dimensional members of that dimensional level;

execute, by one or more computing devices, a query for each of the second plurality of dimensional levels using the dimensional members of each of the second plurality of dimensional levels resulting in a set of data values for each of the second plurality of dimensional levels, wherein each data value of the set of data values is calculated by at least applying a value of one of the dimensional members of one dimensional level of the second plurality of dimensional levels to the data value of interest, wherein a subset of the second plurality of dimensional levels outside the context and the drill path is used for the execution of the query;

calculate a variance of the set of data values generated for each dimensional level of the second plurality of dimensional levels;

identify one of the dimensional levels of the second plurality of dimensional levels as having a greatest variance of the calculated variances;

select the identified dimensional level as the driving factor;

generate a new data value of interest having a new context that comprises an intersection of the first plurality of dimensions and a dimension of the driving factor;

select the new data value of interest;

remove levels involved with the new data value of interest from the second plurality of dimensional levels;

determine a new driving factor; and

add the new driving factor to the context.

8. The system of claim 5, wherein the business intelligence server device is further configured to enumerate the second plurality of dimensional levels outside the context and outside the drill path in a list.