Partitioning is one of the most powerful features for data warehouses in Oracle databases. In this blog post, I explain how it can be used for the physical design of star schemas. What is the recommended partitioning strategy for a star schema, and what are the advantages of partitioning?
Category Archives: Dimensional Modeling
Star Schema Design in Oracle: Fundamentals
What are the design rules for good performance in a star schema in an Oracle database? This blog post series introduces some recommendations for the physical database design. This first post is about constraints and indexes.
Star Schema Optimization in Autonomous Data Warehouse Cloud
Oracle Autonomous Data Warehouse Cloud does not allow to create indexes. Is this a problem for star schemas because no Star Transformation can be used? Or are the required bitmap indexes automatically created? A look under the hood of ADWC.
My Sessions at DOAG 2017
This year, I had the opportunity to present three sessions at the DOAG conference in Nuremberg – one on each conference day. Here a short summary of the sessions and links to the downloads.
Analytic Views: Powerful Oracle 12.2 Feature for Business Intelligence
Analytic Views are one of the main features for Business Intelligence introduced with Oracle 12c Release 2. They provide a lot of new possibilities for analytical queries on star schemas. This blog post gives an overview of the new functionality.
Attribute Dimensions and Hierarchies in Oracle 12.2
Oracle 12c Release 2 introduced Analytic Views, a new set of metadata objects that are very useful for Data Warehouses and Business Intelligence applications. In the first blog post about this new feature I will have a detailed look at two of the new object types: Attribute Dimensions and Hierarchies.
Derived Measures and Virtual Columns
Virtual Columns in Oracle are a convenient feature to implement derived measures of a fact table. But in combination with Oracle Database In-Memory, it has an impact on performance – unless you use Oracle 12c Release 2.
Loading Dimensions from a Data Vault Model
Loading data into a Data Vault Model is easy. But how complex is it to extract the data from a Data Vault into a dimensional Data Mart? A Point in Time (PIT) table helps to load Dimensions from a Hub with multiple Satellites.
Reference Partitioning on Star Schema: A Good Idea?
In my last blog post, I noticed that surrogate keys on a time dimension cannot be combined with partitioned fact tables in a star schema. Actually, there is a way to do that: With reference partitioning it is possible to implement monthly partitioned fact tables based on a surrogate time key. But is this a good idea?
Time Dimension Keys and Partitioning
In the physical design of a star schema, it is recommended to use surrogate keys for the primary keys of dimension tables. Except for the time dimension. In Oracle data warehouses it is common to use a DATE column as the primary key. The reason for this decision and several alternative solutions are described in this post.