You are on page 1of 37

Microsoft SQL Server - Related Courses

• 20761C: Querying Data with Transact-SQL


• 20762C: Developing SQL Databases
• 20764C: Administering a SQL Database Infrastructure
• 20765C: Provisioning SQL Databases
• 20767C: Implementing a SQL Data Warehouse (DW, SSIS, DQS, MDS)
• 20768C: Developing SQL Data Models (SSAS)
• 20778B: Analyzing Data with Power BI
• 20779B: Analyzing Data with Excel
• 20465: Designing Solutions for SQL Server (mostly high availability)
• 20467: Designing Business Intelligence Solutions for Microsoft SQL
Server
• 10985C: Introduction to SQL Databases (PK-FK relationships and
normalization)
• 10986C: Updating Your Skills to SQL Server 2016
• 10987C: Performance Tuning and Optimizing SQL databases
• 10988C: Managing SQL Business Intelligence Operations
• 10990C: Analyzing Data with SQL Server Reporting Services (SSRS)

See the full line of Microsoft Training and Certification resources at:
http://www.microsoft.com/learning/
June 12, 2018

The Absolute Basics of Developing


SSAS Multidimensional and Tabular
Data Models

by George SQUILLACE
“SQL” is in the name!
CuriousGeo@e-Squillace.com
George SQUILLACE
“SQL” is in the name!
 Husband & Dad
 Twenty Five Years of New Horizons service
(June 2016) For fun: (no particular order)
Movies
 ~30 years in Information Technology
Reading
 Microsoft Certified Trainer (MCT) since 1997 Bicycling (road / city / trail)
 99 times in the exam booth! Travel
Music / Acoustic Guitar
 Microsoft SQL Server certifications Computers
 MCSE: Business Intelligence Aircraft
 MCSE: Data Platform
 MCSA
 Microsoft Virtual Academy presenter
 SQLSaturday presenter
 Certified in every version of SQL Server since SQL 2000
 Certified in every version of Windows Server from
NT 4.0 – Windows Server 2008

+ George.Squillace@NHLS.com
" www.e-Squillace.com
Agenda Highlights
What’s it all about?
Demo: Interactive Analysis
SSAS Installation
Multidimensional Models
The B.I. Semantic Model BISM
Tabular Models
Resources
A little bit on B.I.
 It’s tough running an organization.
 To run an organization I have to make
decisions.
 To make good decisions I need good
information.
 Remember, just because I have good
information doesn’t mean I will make a good
decision.
B.I. Landscape
Data model – the “magic” of
Data sources are the data models comes from
optimized for small, a copy of the source data also
random, reads & stored in the model, but in a
writes, which is why proprietary, efficient way, and
those databases are from the aggregations already
normalized and why calculated and stored in the
queries contain model
many joins.

Data warehouse
- Combined,
cleansed,
standardized data
along with
historical data.

Optimized for
large, sequential
reads.
What is SSAS and Why Do I Care?

 The SQL Server Analysis Services (SSAS) service hosts


databases containing Data Models
 SSAS enables rapid, interactive analysis of large volumes of
data and Data Mining capabilities
 SSAS is an “Instance Feature” of Microsoft SQL Server
 The Data Models are accessible by a variety of clients
 A typical SSAS client is an Excel Pivot Table
 SSRS and Power BI are other SSAS client programs
 Two Available Data Model Types
 Multidimensional data model
 Tabular data model (introduced SQL Server 2012)
Data Analysis
Demonstration
SSAS Installation
SSAS is a Product “Feature”

 SSAS is an Instance Feature (as opposed to a Shared Feature)


Instance Configuration
Installation Modes
Multidimensional Models
Concepts: Measures and Dimensions

 Measures

 Dimensions
Development Tool – Visual Studio

 a.k.a. “SQL Server Data Tools” (SSDT)


Multidimensional Model
Development Process
Data Source Definition

 The data source for a multidimensional model is a


data warehouse
 There are generally four inputs to a data source
definition:
 Data Provider (OLE-DB)
 Server name
 Authentication details (Windows or SQL)
 Database name
Data Warehouse Structure
Data Source View Definition

 Identify tables and/or views from the data source to


be included in the data model
 One or more fact tables
 One or more dimension tables
 “Beautification”
 Alias table names
 Alias column names
 Create calculated columns
Cube Wizard

 General approach
 Bottom Up
 Top Down
 Measure Group table selection
 Measure selection
 Dimension selection
Cube Wizard
Demonstration
Subsequent Multidimensional Model
Development Tasks
 Refine Dimensions
 Assign Time intelligence to the date dimension
 Add Attributes to Dimensions
 Create Hierarchies and Levels and Attribute Relationships
 Define Measure Group and Measure Properties
 Partitions
 Aggregations
 Define Other Cube Properties
 Calculations tab
 Named Sets
 Calculated Members
 KPIs
 Actions
 Perspectives
 Translations
Deployment

 Deployment =
 XML file validation
 File copy to an existing SSAS server installed in
Multidimensional and Data Mining mode
 Create and store “Aggregations”
BI Semantic
Model - BISM
What is BISM?
Tabular Models
What is a Tabular Model?

 Based off of Microsoft’s acquired “Vertipaq” /


xVelocity in-memory, column-oriented, hyper-
compressed values engine
 Originally available in Power Pivot in Excel (personal
data model)
 Later became available in SSAS as an installation
mode
 The Power BI data model is Tabular
Tabular Model
Development Process

 Dev Server and Workspace Database


 Create a Tabular project
 Import Data
 Create and Customize Relationships
 Data Modeling tasks
 Assign data types
 Create Custom Columns
 Create Custom Measures and KPIs
Workspace Database

Must have an already existing instance of SSAS


installed in Tabular mode
The database name is a concatenation of a GUID, the
user name, and the SSDT project name
Tabular Model
Demonstration
Measures and Calculated Columns
- DAX
Relationships

 FK >> PK relationships within databases are


automatically detected
 When importing data from disparate sources (i.e.,
DW and a text file) relationships must be manually
created
Deployment

 Acquire a snapshot of the data


 Compress the column-oriented values
 Host them in memory
 Deployment creates a unique Tabular database in the
Tabular SSAS instance
Debate:
Multidimensional or Tabular?
Resources
References & Resources
 Books
 SQL Server 2008 MDX Step by Step
 SQL Server 2012 Analysis Services - The BISM Tabular Model
 There are lots of DAX-specific books
 Software
 Tools for DAX and Tabular Developers
 Videos
 Craig Utley’s videos on Multidimensional model development
(LearnMSBI)
 PragmaticWorks recorded hour-long webinars
 Blogs/Websites
 SQLBI.com (Marco Russo and Alberto Ferrari)
 Chris Webb’s blog
 MSSQLTips.com
 Blue Granite
 SQLServerCentral.com
 Training
 20768 Developing SQL Data Models
Thank you,
and thank our sponsor!

You might also like