www.sqlbi.

com
Who
• BI Experts and Consultants
• Founders of www.sqlbi.com
o Problem Solving
o Complex Project Assistance
o Data Warehouse Assessments and Development
o Courses, Trainings and Workshops
• Book Writers
• Microsoft Business Intelligence Partners
• SSAS Maestros – MVP – MCP

a brand of

Brief History of SQL Some Facts About SSAS

Year Release Number
1993 SQL Server 4.2 • 13 years of life (1999-2012)
1995 SQL Server 6.0
1996 SQL Server 6.5
• OLAP
1998 SQL Server 7.0 o Multidimensional Data Modeling
1999 SQL Server 7.0 OLAP First release of SSAS
o MDX Language / MDX Script
2000 SQL Server 2000 SSAS 2000 – MDX Language
2003 SQL Server 2000 64 bits • Unified Data Modeling
2005 SQL Server 2005 SSAS 2005 –UDM – MDX Script o Metadata layer on top of the DWH
2008 SQL Server 2008 SSAS 2008 – Few improvements
• Huge customer base
2010 SQL Server 2008 R2 SSAS 2008R2 – Very Few improvements

2010 PowerPivot for Excel First introduction of the VertiPaq Engine

2012 SQL Server 2012 BISM – Multidimensional – Tabular

Analysis Services 2012 BI Semantic Model: Architecture

Third-party Reporting SharePoint
• UDM
Power Excel
applications Services PowerPivot Insights
View

o Mature technology
o Few improvements
• xVelocity in-memory engine (VertiPaq)
o Not an evolution of OLAP
o Brand new SSAS Engine
o Programmed and queried with DAX
o New modeling experience (Tabular)
• UDM + Tabular = BISM
Databases LOB Applications Files OData Feeds Cloud Services

1
BI Semantic Model Flavours of BISM
What about existing Applications?
• Multidimensional
o SQL Server Analysis Services
• Tabular
o SQL Server Analysis Services
Existing Existing New
applications applications applications o PowerPivot for SharePoint
Based on Unified Every UDM becomes a BI New technology options
Dimensional Model Semantic Model
o PowerPivot for Excel 2010

Agenda Analysis Services Server Mode

• Create a Tabular model • Choose server mode of Analysis Services
DeploymentMode setting in msmdsrv.ini
o Simple initial model
o 0 – Multidimensional and Data Mining (default)
o Count and Distinct Count measures
o 1 – SharePoint
o New Customers, Returning Customers
o 2 – Tabular
o ABC Analysis
o Security • Different icons in Object Explorer
o Deployment • Same code, different behavior
o Querying
• Choice at the INSTANCE level
• Two SSAS instances needed to use both

Sample Project Key Points

• Prepare development environment • Servers
o Workspace Server
• Implement simple model and measures
o Deployment Server
• Some DAX Patters,
• Project Properties
• Security o Deployment Options
o Deployment Server Edition
• Model Properties
o Data Backup
o Workspace Retention

2
Workspace Database Simple cube
• SSDT always works with the Worskpace DB • Load data from views
o Data is not inside SSDT • Workspace database
o It is stored in the Workspace Database
o Each model update is reflected on the server • Calculated columns

• Workspace databases • Hierarchies to simplify navigation

always use UserName • Measures
and GUID
o Can be safely deleted
when not in use

Count, DistinctCount New Customers

• Count customers/Product NewCustomers:= CALCULATE(
COUNTROWS( Customers ),
FILTER(
• Distinct Count fo buying customers Customers,
Users never bought
anything
CALCULATE(
• New customers or returing customers? COUNTROWS( Sales ),
FILTER(
ALL( Calendar ),
[DateKey] <= MIN ( [DateKey] )
)
) = 0
&&
CALCULATE(
COUNTROWS( Sales ) Users who bought
) > 0 something now
)
)

ABC / Pareto ABC in Excel

• What is ABC analysis?
• Clustering technique
• Helps focusing on Class Percentage
company’s core A (some products) 70%
business B (few products) 20%
C 10%
GRAND TOTAL 100%

3
ABC in SQL ABC in MDX
SUM( Amount)
GROUP BY Product MDX query
INTO #temp Process Sales cube on Sales cube
to extract ABC Class
SQL

OLAP OLAP
UPDATE #temp
SET Percentage =
RunningTotal on Product
Amount / SUM( Amount )

SQL
Process Product UPDATE product
Dimension SET Class = A/B/C
UPDATE product • Update Sales aggregations • from MDX query
SET Class = A/B/C (from (ABC flexible attribute) • Requires ETL SQL
or Linked
#temp) OLAP Server

SQL

Add security Conclusions

• Role based security • Simpler data model
• Similar to SSAS Multidimensional • Don’t use dimensional modeling
• No cell security • Calculated columns – Mesaures
• No differences between dimensions and • DAX changes the way we create data models
maesure group(s)
• Speed is a modeling tool
• Table level security by using DAX

And now? Link

• Learn DAX • SQLBI website
http://www.sqlbi.com
o Really… learn it!
o Is not so easy as it might seem at first sight • Blog by Alberto Ferrari
http://sqlblog.com/blogs/alberto_ferrari/default.aspx

• Blog by Marco Russo

http://sqlblog.com/blogs/marco_russo/default.aspx

• Blog by Jeffrey Wang

http://mdxdax.blogspot.it/

4
 Oh, just another thing…

Are you sure to get the best out of

Thank you! your analytical environment?

Ask to SQLBI, we offer many services to help you:

Check daily our new articles on

www.sqlbi.com

Consulting Assessment Outsourcing Technical

Fellowship

Find out more on

www.sqlbi.com/consulting