Analysis of OLAP data to determine user-relevant information

• US 7,752,162 B2
• Filed: 06/02/2006
• Issued: 07/06/2010
• Est. Priority Date: 06/02/2006
• Status: Active Grant

• Current Assignee: Twitter Inc (X Holdings Corp.)
• Original Assignee: International Business Machines Corporation
• Inventors: Chenthamarakshan, Vijii Enara, Banavar, Guruduth Somasekhara, Kummamuru, Krishna, Deshpande, Prasad Manikarao
• Primary Examiner(s): Ali; Mohammad
• Assistant Examiner(s): Hocker; John P
• Application Number:
  US11/421,864
• Publication Number:
  US 20070282789A1
• Time in Patent Office: 1,495 Days
• Field of Search: 707/1, 707/100, 707/600, 707/602, 705/7
• US Class Current: 707/600
• CPC Class Codes: G06F 16/283 Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP

Abstract

The analysis of OLAP data to determine user-relevant information firstly generates a set of queries based on said preferences. Each query is evaluated sequentially against the OLAP data to give a query result. For each evaluated query in turn, it is determined whether said result is relevant to the user on the basis of conditions derived from the user preferences. An output results set is formed consisting of the relevant results. Further, if a previous query result containing a common measure was determined not to be relevant, then a subsequent query can be omitted from evaluation.

Claims

1. A computer-implemented method for analyzing on-line analytical processing (OLAP) data to determine user-relevant information, said method comprising:

inputting a set of user preferences expressed in a user preference specification language to a computer,

wherein each user preference includes a subset of dimensions, measures, conditions on measures, a drill down specification that specifies a dimension upon which to drill down, and conditions upon which said drill down occurs;

generating, by said computer, a set of queries, relevant to said user, based on said set of user preferences, said generating a set of queries, relevant to said user, comprising;

generating queries sequentially, said generating queries sequentially comprising;

generating a single initial query based on said subset of dimensions, measures, and conditions on measures of each user preference, and

further sequentially generating queries based on drilling down according to said drill down specification, while said conditions upon which said drill down occurs are satisfied;

or generating a set of queries, one for each combination of cube dimensions that are of interest;

determining, by said computer, how often each of said subset of dimensions and said measures, corresponding to said each user preference, is accessed by said user for a query;

evaluating, by said computer, each of said queries against said OLAP data to give a query result;

determining, by said computer, for each of said evaluated queries, whether said query result is relevant to said user, based on said conditions of said each user preference;

if said query result is not relevant, then preventing, by said computer, subsequent queries from being generated for said evaluating, based on said subset of dimensions, said measures, said conditions on measures, and said drill down specification for said each user preference corresponding to said query result that is not relevant; and

if said query result is relevant, then adding, by said computer, said query result to relevant query results that are to be outputted;

based upon said determination, by said computer, of how often said subset of dimensions and said measures are accessed by said user, and said user selected conditions for tagging said queries and query results as interesting, and for drilling down on a selected dimension,

modifying, by said computer, said set of user preferences expressed in said user specification language to include said subset of dimensions and said measures that are often accessed by said user, said user selected condition for tagging said queries and query results as interesting, and said selected condition for drilling down; and

outputting, by said computer, said relevant query results to said user.

2. The method of claim 1, wherein said conditions on measures comprise any of Boolean conditions, thresholds for outliers, statistical conditions, trends, and comparisons.

3. The method of claim 1, wherein said conditions on measures are represented as a dependency graph.

9. The method of claim 1, wherein if a previous query result, including a common measure, is determined not to be relevant, then omitting said evaluating of a subsequent query including said common measure.

4. A computer system for analyzing on-line analytical processing (OLAP) data to determine user-relevant information, said system comprising:

a memory that stores an inputted set of user preferences expressed in a user preference specification language,

wherein each user preference includes a subset of dimensions, measures, conditions on measures, and a drill down specification that specifies a dimension upon which to drill down and conditions upon which said drill down occurs;

a processor that;

generates a set of queries, relevant to said user, based on said set of user preferences, wherein generating a set of queries, relevant to said user, comprises;

generating queries sequentially, said generating queries sequentially comprising;

generating a single initial query based on said subset of dimensions, measures, and conditions on measures of each user preference, and

further sequentially generating queries based on drilling down according to said drill down specification, while said conditions upon which said drill down occurs are satisfied;

or generating a set of queries, one for each combination of cube dimensions that are of interest;

determines how often each of said subset of dimensions and said measures, corresponding to said each user preference, is accessed by said user for a query,

evaluates each of said queries against said OLAP data to give a query result;

determines, for each of said evaluated queries, whether said query result is relevant to said user, based on said conditions of said each user preference;

if said query result is not relevant, then prevents subsequent queries from being generated for evaluations, based on said subset of dimensions, said measures, said conditions on measures, and said drill down specification for said each user preference corresponding to said query result that is not relevant;

if said query result is relevant, then adds said query result to relevant query results that are to be outputted;

based upon determinations of how often said subset of dimensions and said measures are accessed by said user, and said user selected conditions for tagging said queries and query results as interesting, and for drilling down on a selected dimension, modifying said set of user preferences expressed in said user specification language to include said subset of dimensions and said measures that are often accessed by said user, said user selected condition for tagging said queries and query results as interesting, and said selected condition for drilling down; and

outputs said relevant query results to said user.

5. The computer system of claim 4, wherein said conditions on measures comprise any of Boolean conditions, thresholds for outliers, statistical conditions, trends, and comparisons.

6. The system of claim 5, wherein said conditions on measures are represented as a dependency graph.

10. The system of claim 4, wherein if a previous query result, including a common measure, is determined not to be relevant, then said processor omits said evaluating of a subsequent query including said common measure.

7. A computer program storage medium readable by a computer, tangibly embodying a computer program of instructions executable by said computer to perform a method for analyzing on-line analytical processing (OLAP) data to determine user-relevant information, said method comprising:

inputting a set of user preferences expressed in a user preference specification language,

wherein each user preference includes a subset of dimensions, measures, conditions on measures, and a drill down specification that specifies a dimension upon which to drill down and conditions upon which said drill down occurs;

generating a set of queries, relevant to said user, based on said set of user preferences, said generating a set of queries, relevant to said user, comprising;

generating queries sequentially, said generating queries sequentially comprising;

generating a single initial query based on said subset of dimensions, measures, and conditions on measures of each user preference, and further sequentially generating queries based on drilling down according to said drill down specification, while said conditions upon which said drill down occurs are satisfied;

or generating a set of queries, one for each combination of cube dimensions that are of interest;

determining how often each of said subset of dimensions and said measures, corresponding to said each user preference, is accessed by said user for a query;

evaluating each of said queries against said OLAP data to give a query result;

determining, for each of said evaluated queries, whether said query result is relevant to said user, based on said conditions of said each user preference;

if said query result is not relevant, then preventing subsequent queries from being generated for said evaluating, based on said subset of dimensions, said measures, said conditions on measures, and said drill down specification for said each user preference corresponding to said query result that is not relevant; and

if said query result is relevant, then adding said query result to relevant query results that are to be outputted;

based upon said determination of how often said subset of dimensions and said measures are accessed by said user, and said user selected conditions for tagging said queries and query results as interesting, and for drilling down on a selected dimension, modifying, said set of user preferences expressed in said user specification language to include said subset of dimensions and said measures that are often accessed by said user, said user selected condition for tagging said queries and query results as interesting, and said selected condition for drilling down; and

outputting said relevant query results to said user.

8. The computer program storage medium of claim 7, wherein said conditions on measures comprise any of Boolean conditions, thresholds for outliers, statistical conditions, trends, and comparisons.