Abhinav Gupta, 41Los Gatos, CA

6775681, Aug 10, 2004

Feb 26, 2002

10/084597

Abhinav Gupta - Palo Alto CA
Andrew Witkowski - Foster City CA

Oracle International Corporation - Redwood Shores CA

G06F 1730

707102, 707 3

Methods are provided for efficiently evaluating database queries that include a grouping sets specification, by reducing the grouping sets specification to one or more combinations of groupings that are computable with simple GROUP BY clauses, with or without a rollup operator, wherein the GROUP BY clauses and rollup operators are efficient operations that can be computed with a single data sort. Computational costs are further minimized by computing a grouping from records generated from computation of a previously computed grouping. An optimization algorithm is described which determines the respective grouping combinations to which the grouping sets specification is reduced and the respective GROUP BY clauses, to minimize computational costs. Furthermore, an intermediate grouping, that is, a grouping that is not specified by the query, is included in a query evaluation plan in some instances, to compute specified groupings when doing so reduces computational costs.

6775682, Aug 10, 2004

Feb 26, 2002

10/084642

Abhinav Gupta - Palo Alto CA
Andrew Witkowski - Foster Ctiy CA

Oracle International Corporation - Redwood Shores CA

G06F 1730

707102

Methods are provided for efficiently evaluating database queries including a rollup operator and a distinct aggregate function. Using a sequence of sorts, duplicate record elimination performed on previous sorts at lower, or finer, levels of the rollup operator is taken advantage of by performing subsequent sorts on the preceding sort. Hence, when moving from one rollup level to the next higher level, there are fewer data records to sort with respect to the relevant grouping columns for that level, and thus also fewer duplicate data records to eliminate for purposes of computing the distinct aggregate. Using parallel evaluation, processing of aggregate functions is split among different processing slaves, and the measure of an aggregate function is included as a partitioning key when sending data from one data flow operation to the next data flow operation of a query execution plan. Using parallel evaluation for a query that includes two or more aggregate functions, a measure code corresponding with each aggregate function and associated measure values are included as partitioning keys for enhanced load balancing and parallelization.

6957225, Oct 18, 2005

May 30, 2002

10/160585

Mohamed Zait - Sunnyvale CA, US
Abhinav Gupta - Palo Alto CA, US
Benoit Dageville - Foster City CA, US

Oracle International Corporation - Redwood Shores CA

G06F017/30

707102, 707 2, 707100

Methods are provided for automatically discovering correlations between values in columns of tables. A set of significantly correlated columns is identified by identifying correlated columns, and by determining the significance of the correlation between the correlated columns from one or more tables. If the correlated columns are considered significantly correlated, a correlation table is constructed that includes records representing distinct combinations of values corresponding to the correlated columns. Embodiments include methods for identifying correlated columns, for determining the significance of the correlation between the correlated columns, and for using the resultant correlation table to enhance performance of a query execution process. One particular embodiment provides for using a correlation table for partition pruning a partitioned table, with respect to a query execution plan.

7035843, Apr 25, 2006

Feb 15, 2002

10/077828

Srikanth Bellamkonda - Sunnyvale CA, US
Abhinav Gupta - Palo Alto CA, US
Andrew Witkowski - Foster City CA, US

Oracle International Corporation - Redwood Shores CA

G06F 17/30

707 3, 707100

Methods are provided for efficiently evaluating database queries including multiple rollup operators. With the computation of grouping identifiers to distinguish grouping levels of each rollup operator, evaluation of database queries that include concatenation of rollup operators includes nesting of the rollup operators and includes the grouping identifiers as sorting keys for subsequent processing. Furthermore, to optimize the query evaluation process, the order in which the rollup operators are computed can be determined based on cardinalities that estimate the number of records generated by each rollup operator, and parallel evaluation can be utilized by partitioning data records between rollup operator processing stages based on the grouping keys associated with the rollup operators that are not being processed at the next stage. If the query includes an algebraic aggregate function, the algebraic function is represented as distributive functions, which are propagated through the evaluation stages and used to compute the algebraic function at the final rollup stage.

7111020, Sep 19, 2006

Mar 26, 2002

10/107106

Abhinav Gupta - Palo Alto CA, US
Andrew Witkowski - Foster City CA, US

Oracle International Corporation - Redwood Shores CA

G06F 17/30

707201, 707 2

Techniques are provided for improving efficiency of database systems, and in particular, to refreshing materialized views maintained by database systems and rewriting queries to access the materialized views. According to the approaches, a ranked materialized view is incrementally refreshed, and during the incremental refresh operation, rows in the partitions of the materialized view are ranked within the partitions.

7177855, Feb 13, 2007

Nov 6, 2003

10/704192

Andrew Witkowski - Foster City CA, US
Srikanth Bellamkonda - Sunnyvale CA, US
Tolga Bozkaya - Belmont CA, US
Abhinav Gupta - Santa Clara CA, US
Nathan Folkert - Palo Alto CA, US
Sankar Subramanian - Cupertino CA, US

Oracle International Corporation - Redwood Shores CA

G06F 17/30
G06F 7/00

707 2, 707 3, 707 4, 707 5

Described herein are optimizations and execution strategies for spreadsheet extensions to SQL. The partitioning of data, as specified in a spreadsheet clause, provides a way to parallelize the computation of spreadsheet and to provide and improve scalability. Even if the partitioning is not explicitly specified in the spreadsheet clause, the database optimizer can automatically infer the partitioning in some cases. Efficient hash based access structures on relations can be used for symbolic array addressing, enabling fast computation of formulas. When rewriting SQL statements, formulas whose results are not referenced in outer blocks can be removed from the spreadsheet clause, thus removing unnecessary computations. The predicates from other query blocks can be moved inside query blocks with spreadsheets clauses, thus considerably reducing the amount of data to be processed. Conditions for validity of this transformation are given.

7356542, Apr 8, 2008

Mar 8, 2004

10/796217

Abhinav Gupta - Santa Clara CA, US
Lei Sheng - Foster City CA, US
Sankar Subramanian - Cupertino CA, US
Nathan Folkert - San Francisco CA, US

Oracle International Corporation - Redwood Shores CA

G06F 17/00
G06F 7/00

707102, 707103 Y

A method and apparatus for densification of data is provided. A method and apparatus for including a construct for partitioning data within a DML statement is provided. The construct for partitioning data may not necessarily perform a densification of data, and the densification of data may not necessarily include a construct for partitioning data. In an embodiment, the syntax of an OUTER JOIN is extended to include a PARTITION BY construct that may be used for data densification.

7761403, Jul 20, 2010

Nov 6, 2003

10/703412

Andrew Witkowski - Foster City CA, US
Srikanth Bellamkonda - Sunnyvale CA, US
Tolga Bozkaya - Belmont CA, US
Abhinav Gupta - Santa Clara CA, US
Nathan Folkert - Palo Alto CA, US
Sankar Subramanian - Cupertino CA, US

Oracle International Corporation - Redwood Shores CA

G06F 17/30

707 4, 707 2, 707 7, 707 8, 707 9, 707 10

Described herein are optimizations and execution strategies for spreadsheet extensions to SQL. The partitioning of data, as specified in a spreadsheet clause, provides a way to parallelize the computation of spreadsheet and to provide and improve scalability. Even if the partitioning is not explicitly specified in the spreadsheet clause, the database optimizer can automatically infer the partitioning in some cases. Efficient hash based access structures on relations can be used for symbolic array addressing, enabling fast computation of formulas. When rewriting SQL statements, formulas whose results are not referenced in outer blocks can be removed from the spreadsheet clause, thus removing unnecessary computations. The predicates from other query blocks can be moved inside query blocks with spreadsheets clauses, thus considerably reducing the amount of data to be processed. Conditions for validity of this transformation are given.