SteelWheels
Walk through of SteelWheels domain ..
Workshop - Overview of SteelWheels Domain
Creating effective business intelligence solutions requires more than just connecting to databases—it demands a semantic layer that translates technical database structures into business-friendly concepts that analysts and report builders can understand. Pentaho Metadata Editor enables you to build this critical abstraction layer, organizing raw tables and columns into logical business models with meaningful names, relationships, and categories that mirror how your organization thinks about data.
In this hands-on workshop, you'll explore the SteelWheels domain- a comprehensive metadata model built on Pentaho's sample database that represents a fictional toy company's complete operations. By examining this fully-realized example spanning customers, products, orders, employees, and financial data across 13 interconnected tables, you'll learn the fundamental architecture of metadata domains and how each layer builds upon the previous one to create an intuitive, business-focused data access layer.
What You'll Accomplish:
Import and explore an existing metadata domain using the XMI file format
Examine the Physical Layer including database connections, tables, and column definitions
Navigate Business Models and understand how they organize tables for specific analytical purposes
Explore Business Tables and Business Columns with their enhanced metadata properties
Understand Relationships and how they define connections between tables for accurate query generation
Discover Business Views and Categories that expose the semantic layer to end users
Learn the complete metadata architecture from physical structures to business-friendly presentations
By the end of this workshop, you'll understand how Pentaho Metadata Editor transforms raw database schemas into organized, documented, business-aligned data models. You'll see how physical tables become business tables with friendly names, how relationships enable cross-table analysis, and how categories group related fields into logical collections that match business thinking. This foundational knowledge prepares you to build your own metadata domains that empower business users to create reports and analyses without needing to understand underlying database complexity.
Prerequisites: Basic understanding of relational database concepts (tables, columns, relationships); Pentaho Metadata Editor installed and configured; Access to SteelWheels sample database
Estimated Time: 45 minutes
Follow the guide below to understand how a Domain is defined:
SteelWheels Domain
The following steps show you how to use an example metadata model in Pentaho Metadata Editor to create an Interactive Reports.
Start Metadata Editor:
Import (open) an existing metadata domain:
From the menu, select: File > Import from XMI File
Navigate to:
Double-click: steel-wheels.xmi.
In the Save Model dialog, type: SteelWheels.
Click OK.
Note a domain can only have 1 database connection.
Physical Layer
The physical layer of a Pentaho domain encompasses connections, physical tables and physical columns. These objects represent the database(s) you are trying to model and enrich with metadata. The physical layer is not considered part of the business model, because not all connections defined in the physical layer will be used in every business model.
Once imported you will see that:
Physical Layer includes 13 tables.
In the left pane, expand Connections > SampleData.
Expand Products and notice the list of columns.
Double-click on Product Line.
A Properties Editor opens, which displays the metadata properties for each object (prefix):
PT - Physical Table
PC - Physical Column
BT - Business Tables
BC - Business Columns
CT - Categories
Scroll through the Base settings - Parent Concepts.
Notice most of the Base metadata settings can be overridden. The property - Parent Concept - will be inherited in the Business Models and Business Views Layers.
Expand > Orderdetails table.
Double-click on Total.
If you have a requirement to create dimensions / measures for Interactive Reports that desn't exist in the Physical tables, then you can easily create the column based on existing columns.
Scroll through the Base settings.
Notice the column is based on the calculation:
[QUANTITYORDERED]*[PRICEEACH]
The columns are referenced with [COLUMN]
This can be aggregated to: Sum, Average, Minimum & Maximum.
Business Models
There are three Business Models within this metadata domain. Each model uses one or more tables from the physical layer.
Collapse Connections and Expand: Business Models.
Expand: Orders Business Model.
The Orders Business Model includes five Business Tables:
Customer W Ter
Orders
Orderdetails
Products
Payments
The Business Model Layer is where you overide and set the Parent Concept to be inherited in the Business View Layer.
Expand: Orderdetails Business Table.
Double-click: Price Sold.
Notice a 'Mask for Number or Date' has been added and the value set as a Parent Concept. The Mask - $#,##0.00 format - will be inherited in the Business View Layer.
Relationships
Once you have all your Business Tables created, you will need to create relationships between the tables, so that the query generators and SQL generators that work with Pentaho metadata can create the data queries correctly.
This is very much like drawing a relational diagram to show primary and foreign key relationships. Although relational links are not the only relationships that can be modelled. You can create a relationship between any two tables, link any two columns between them and dictate what the relationship is (one to many, many to many, etc.).
The important pieces of information to know before you try to create a relationship is:
what two Business Tables would you like to associate with this relationship?
what columns in the business tables identify the relationship?
and what kind of relationship is it - one to one, one to many, many to one, etc?
Expand: Business Models > Orders > Relationships
Double-click:
The relationship defines the Inner JOIN with the constraint 1 order has 1 payment.
The SQL is generated from references the unique keys in the Tables using the following metadata syntax:
Business Views
Business Views acts as 'buckets - categories' that enable you to define the sematic layer that is exposed to the end user. Each Category can contain any column(s) that have been defined in the Business tables.
Expand: Business View > Orders
Highlight Orders category.
Right-mouse click and select the option: Manage Categories.
The Panel enables you to associated the Business Column(s) with the Category.
Was this helpful?