MDX Essentails
Paul Turley
Mentor, SQL Server MVP
October 11-14, Seattle, WA
Introduction
Paul Turley
Mentor, SQL Server MVP
SqlServerBiBlog.com
1
What Can You Do with a Cube?
> Aggregate very large
> Destroy anything volumes of data
in its path
> Present browse-able
business information for
self-service reporting
> Assimilate entire
> Create high-value
civilizations
business reports that
render in a fraction of
the time of a
relational data source
> Create a mega race
of neo-humanoid > Encapsulate complex
androids with a business rules into
single collective predefined hierarchies,
consciousness calculations, business
measures and KPIs
2
The Business Data Continuum
Relational Data
Warehouse
Data Consolidation OLAP Cubes
& Transformation
(ETL)
Operational
Databases Reports, Charts,
Dashboards & Scorecards
3
MDX
M ulti
D imensional
eX pressions
• A query language developed & used by Microsoft
& other vendors for OLAP data products
• Part of the OLE DB for OLAP 1997 specification held
by Microsoft (not ANSI or ISO)
• Specification was written mainly by Mosha
Pasumansky from Microsoft, originally from
Panorama
4
Why Can’t I Use SQL?
Multidimensional
storage is a
fundamental paradigm
shift from relational
storage
MDX is optimized for
dimensional hierarchies
& not table/index
scans
5
Dimensional Data Warehouse Design
Customer Date
Dimension Dimension
Product Sales Employee
Dimension
Dimension Fact
Vendor Geography
Dimension Dimension
6
Dimensions
Dimension > Hierarchy > Level > Member
7
Measures
Organized in
measure groups
Derived from numeric
fields or SQL
calculations
8
Basic Query Syntax
SELECT
< member or set > on < Columns | Axis(0) | 0 >,
< member or set > on < Rows | Axis(1) | 1 >
FROM
< cube or subcube expression >
WHERE
< member or set > ;
SELECT
{ [Sales Amount], [Order Quantity] } on Columns,
[Category].Members on Rows
FROM
[Adventure Works]
WHERE
[CY 2001] ;
9
Filtering
Slicer
SELECT … on Columns, … on Rows
FROM < cube name >
WHERE { [Category].[Bikes], [Category].[Clothing] }
;
Subcube
SELECT … on Columns, … on Rows
FROM
(
SELECT { [Category].[Bikes], [Category].[Clothing] } on 0
FROM < cube name >
)
;
10
Sets
Combine members from same hierarchy using
braces
{ [Year].[2005], [Year].[2006] }
11
Tuples
Combine members from different hierarchies
using parentheses
( [Category].[Bikes], [Year].[2006] )
Add measure to return a value from the tuple:
(Measures.[Sales Amount]
, ( [Category].[Bikes], [Year].[2006] ) )
12
Calculated Members
Calculation added to Measures:
WITH Member [Measures].[Product Percent of Parent] AS
[Measures].[Sales Amount] / ( [Measures].[Sales Amount]
, [Product].[Product Categories].Parent )
, Format_String = "Percent"
Used in Query:
SELECT { [Measures].[Sales Amount]
, [Measures].[Product Percent of Parent] } on Columns
, NON EMPTY [Product Categories].Category.Members on Rows
FROM < cube name > ;
13
Calculated Members
Calculation added to Hierarchy:
WITH Member [Product].[Product Categories].[All].[Category Average] AS
Avg([Product].[Product Categories].[Category])
Used in Query:
SELECT { [Measures].[Sales Amount]
, [Measures].[Product Percent of Parent] } on Columns
, NON EMPTY [Product Categories].Category.AllMembers on Rows
FROM < cube name > ;
14
Member Functions
.CurrentMember Returns the current member of a hierarchy
Typically used in a calculated member expression
.Parent Returns the parent member of a member
Ancestor ( ) Returns a member at a level above a member
in a user hierarchy – by level name or distance
.PrevMember Returns a sibling or cousin one position prior
ParallelPeriod ( ) Returns a member at the same level and
ordinal position of a member, relative to
a specified ancestor
.Item ( ) Returns a specified ordinal member of a set
IsEmpty ( ) Returns True for a tuple containing a value
Is Null Returns True if a member exists
15
Set Functions (1)
.Children Returns set a level below a member
Descendants ( ) Returns set any level below a member
(named level or distance from level)
Head ( ) Returns the first X members of a set
Tail ( ) Returns the last X members of a set
Union ( ) or
Set1 + Set2 Combines 2 sets into one set (Distinct or All)
Intersection ( ) Returns the common members of 2 sets
Except ( ) or
Set1 - Set2 Returns members of 1st set that don't exist in 2nd set
16
Set Functions (2)
Order ( ) Returns a set in the order of a
specified member/measure
TopCount ( ) Returns top X ranking members based on a measure
Hierarchize ( ) Organizes & sorts members of a set by
hierarchal structure. Typically used on Unioned sets .
Filter ( ) Filters a set based on a Boolean expression
Exists ( ) Filters one set by another set from the same dimension
Returns only members of the 1st set, not the 2nd set
NonEmpty ( ) Filters one set by a set from a different dimension
Returns only members of the 1st set, not the 2nd set
17
Set Functions (3)
Sum ( ) Returns a numeric sum for a measure over a set
Aggregate ( ) Like Sum() but uses the measure's AggregateFunction
LastPeriods ( ) Returns a set of previous periods including current
PeriodsToDate ( ) Returns a set of current level members sharing a
common time level ancestor
YTD ( ) Simplified PeriodsToDate( ) with fixed time level
QTD, MTD, WTD - YTD
Generate ( ) Returns a string from a set, concatenating a
member property separated by a specified delimeter
SetToStr ( ) Returns a string from a set.
Delimits full UniqueName reference for each member
separated by commas
18
Crossjoins
Same Dimension
Returns coexisting members
([Product].[Category].[Category], [Product].[Color].[Color])
([Product].[Category].[Category], [Product].[Color].[Red])
Different Dimensions
Returns Cartesian product
([Product].[Color].[Color], [Date].[Calendar Year].[Calendar Year])
([Product].[Color].[Color], [Date].[Calendar Year].[2001])
19
Questions
20
Thank You
Resources
Paul’s Blog……SqlServerBiBlog.com
[email protected]
SQL Server 2008 MDX
Bryan C Smith, Ryan Clay
Microsoft Press
SQL Server 2008
Analysis Services
Scott Cameron
Microsoft Press
21 October 11-14, Seattle, WA
