table.
Oracle8 introduced a further mechanism called a parallel bitmap star join , which uses bitmap indexes on the foreign keys to the dimension tables to speed star joins involving a large number of dimension tables.
Materialized views
Since Oracle8 i , materialized views have provided another means of achieving a significant speedup of query performance. Summary-level information derived from a fact table and grouped along dimension values is stored as a materialized view. Queries that can use this view are directed to the view, transparently to the user and the SQL they submit. Oracle has continued to improve optimizer usage of materialized views with each new release of the database.
Analytic functions
A growing trend in Oracle and other databases is inclusion of SQL-accessible analytic and statistical functions in the database. Oracle first introduced such capabilities in Oracle8 i with the CUBE and ROLLUP functions. Today, the functionality provided also includes ranking functions, windowing aggregate functions, lag and lead functions, reporting aggregate functions, statistical aggregates, linear regression, descriptive statistics, correlations, crosstabs, hypothesis testing, distribution fitting, and Pareto analysis.
OLAP Option
The OLAP Option physically stores dimensionally aware cubes in the Oracle relational database. These cubes are most frequently accessed using SQL, although a Database Performance Features
|
21
Java API is also supported. As of Oracle Database 11 g , Oracle’s optimizer recognizes the levels within these cubes. As a result, any business intelligence tool that submits SQL to an Oracle database can transparently take advantage of the improved performance offered by deployment of this option. Refreshes of the values in these cubes are now maintained similar to refreshing materialized views.
Data Mining Option
Since Oracle9 i , popular data-mining algorithms have been embedded in the database through the Data Mining Option and are exposed through a PL/SQL or Java data-mining API. Data-mining applications that use these algorithms are typically built using Oracle’s DataMiner or using data-mining tools from Oracle partners such as InforSense and SPSS. Algorithms available in the Data Mining Option for Oracle Database 11 g include Naïve Bayes, Associations, Adaptive Bayes Networks, Clustering, Support Vector Machines (SVM), Nonnegative Matrix Factorization (NMF), Decision Trees, and Generalized Linear Models.
Business intelligence tools
Oracle data warehouses are often accessed using business intelligence tools from other popular vendors. However, Oracle’s own tools became more common for such deployment as Oracle grew its offerings through acquisitions. Oracle’s initial offering included Oracle Discoverer and Reports, and these tools remain available in the Application Server or as a standalone Oracle Business Intelligence Standard Edition Suite.
Oracle’s flagship product in this area is Oracle Business Intelligence Enterprise Edition Suite (OBI EE) originally consisting of former Siebel Analytics, including Oracle Answers, Dashboards, Delivers, BI Publisher, and Office Plug-ins. Oracle expanded this offering in OBI EE Plus adding Hyperion components that include Foundation Services, Interactive Reporting, SQR production reporting, Financial Reporting, SmartView for Office, and Web Analysis.
Essbase is available as an option for providing an OLAP cube and functionality independently of the data warehouse database. A subset OBI EE is included in Business Intelligence Standard Edition One, along with the Oracle Standard Edition One database and Oracle Warehouse Builder.
Oracle also offers business intelligence applications that include data models and reporting and analysis with prepopulated business metadata. Flagship applications include Oracle’s Business Intelligence Applications (the former Siebel Business Analytics applications) and Hyperion Financial Performance
Oracle8 introduced a further mechanism called a parallel bitmap star join , which uses bitmap indexes on the foreign keys to the dimension tables to speed star joins involving a large number of dimension tables.
Materialized views
Since Oracle8 i , materialized views have provided another means of achieving a significant speedup of query performance. Summary-level information derived from a fact table and grouped along dimension values is stored as a materialized view. Queries that can use this view are directed to the view, transparently to the user and the SQL they submit. Oracle has continued to improve optimizer usage of materialized views with each new release of the database.
Analytic functions
A growing trend in Oracle and other databases is inclusion of SQL-accessible analytic and statistical functions in the database. Oracle first introduced such capabilities in Oracle8 i with the CUBE and ROLLUP functions. Today, the functionality provided also includes ranking functions, windowing aggregate functions, lag and lead functions, reporting aggregate functions, statistical aggregates, linear regression, descriptive statistics, correlations, crosstabs, hypothesis testing, distribution fitting, and Pareto analysis.
OLAP Option
The OLAP Option physically stores dimensionally aware cubes in the Oracle relational database. These cubes are most frequently accessed using SQL, although a Database Performance Features
|
21
Java API is also supported. As of Oracle Database 11 g , Oracle’s optimizer recognizes the levels within these cubes. As a result, any business intelligence tool that submits SQL to an Oracle database can transparently take advantage of the improved performance offered by deployment of this option. Refreshes of the values in these cubes are now maintained similar to refreshing materialized views.
Data Mining Option
Since Oracle9 i , popular data-mining algorithms have been embedded in the database through the Data Mining Option and are exposed through a PL/SQL or Java data-mining API. Data-mining applications that use these algorithms are typically built using Oracle’s DataMiner or using data-mining tools from Oracle partners such as InforSense and SPSS. Algorithms available in the Data Mining Option for Oracle Database 11 g include Naïve Bayes, Associations, Adaptive Bayes Networks, Clustering, Support Vector Machines (SVM), Nonnegative Matrix Factorization (NMF), Decision Trees, and Generalized Linear Models.
Business intelligence tools
Oracle data warehouses are often accessed using business intelligence tools from other popular vendors. However, Oracle’s own tools became more common for such deployment as Oracle grew its offerings through acquisitions. Oracle’s initial offering included Oracle Discoverer and Reports, and these tools remain available in the Application Server or as a standalone Oracle Business Intelligence Standard Edition Suite.
Oracle’s flagship product in this area is Oracle Business Intelligence Enterprise Edition Suite (OBI EE) originally consisting of former Siebel Analytics, including Oracle Answers, Dashboards, Delivers, BI Publisher, and Office Plug-ins. Oracle expanded this offering in OBI EE Plus adding Hyperion components that include Foundation Services, Interactive Reporting, SQR production reporting, Financial Reporting, SmartView for Office, and Web Analysis.
Essbase is available as an option for providing an OLAP cube and functionality independently of the data warehouse database. A subset OBI EE is included in Business Intelligence Standard Edition One, along with the Oracle Standard Edition One database and Oracle Warehouse Builder.
Oracle also offers business intelligence applications that include data models and reporting and analysis with prepopulated business metadata. Flagship applications include Oracle’s Business Intelligence Applications (the former Siebel Business Analytics applications) and Hyperion Financial Performance
Similar Books
