p. 1 1 Chapter 2 - Video # 9 Measures, Dimensions, and Cubes 1 Chapter 2: Business Intelligence & Data Warehousing with SSAS Course: SQL Server 2008/R2 Analysis Services Course Id: 165 Presented by Scott Whigham
Mar 28, 2016
p. 1
1
Chapter 2 - Video # 9
Measures, Dimensions, and Cubes
1
Chapter 2: Business Intelligence & Data Warehousing with SSASCourse: SQL Server 2008/R2 Analysis ServicesCourse Id: 165Presented by Scott Whigham
p. 2
2
• An MDB is comprised of one or more cubes– Subsets of the database
– Granular at the specific business function/topic
Measures, Dimensions & Cubes
p. 3
3
• The numbers on the face of the cube (called a “cell”) are the measures– Measures are numbers are stored in the MDB in
pre-calculated aggregate form
Measures, Dimensions & Cubes
p. 4
4
• Common measures:– UnitsSold
– SalesAmount
– TaxAmount
– DiscountAmount
– NumberOfDefects
– NumberOfCustomers
– AverageUnitPrice
Measures, Dimensions & Cubes
p. 5
5
• Measures are numbers that your analysts/execs/managers want to know– Measures are answers
• “How many sales did we do?”
• “How many defects have we had?”
• “How much in taxes did we pay?”
Measures, Dimensions & Cubes
p. 6
6
• Dimensions are the context for your measures– Measures are answers
– Dimensions are how you want to view the answer
• “How many sales did we do over each of the past five years?”
• “How many defects have we had for each machine?”
• “How much in taxes did we pay for each salesperson?”
Measures, Dimensions & Cubes
p. 7
7
• Our cube analogy is a great visual for understanding the relationship dimensions and measures have
Measures, Dimensions & Cubes
p. 8
8
• Dimensions can contain hierarchies– A “Time” dimension can roll up from “Day” all the
way to “Year”
Measures, Dimensions & Cubes
p. 9
9
• A Geography dimension might look like this:
• Country, State/Province, City, Postal Code
Measures, Dimensions & Cubes
p. 10
10
• We’ll cover dimensional design in detail in chapters 4 and 5
• For now: back to the cube!
Measures, Dimensions & Cubes
p. 11
11
• A cube is a part of a multidimensional database– It is not the multidimensional database; it is just a
subset of the database
– A cube contains measures and dimensions
– The data stored in a cube is pre-calculated aggregates at various levels of dimensions and their hierarchies
Measures, Dimensions & Cubes
p. 12
12
• “Scott – you keep talking about pre-calculated aggregates over and over… Get to the point!”
Measures, Dimensions & Cubes
p. 13
13
• Okay – so back to our discussion of relational databases– “Groannnnnnnnnnnnnnnn….”
Measures, Dimensions & Cubes
p. 14
14
• Because the cube contains pre-calculated aggregates, theoretically drilldown and drillthrough are extremely fast– It is simply a matter of “pivoting the cube to the
right face”
Measures, Dimensions & Cubes
p. 15
15
• Reports on relational data can perform drilldown and drillthrough but must wade through potentially millions of records with each drill-action!
Measures, Dimensions & Cubes
p. 16
16
• “Wait – did you just imply that I don’t need an MDB; that I can do everything that I can with an MDB with my relational database?”– Why yes, I did!
Measures, Dimensions & Cubes
p. 17
17
• The whole point of an MDB is to make it:– Easier to write analytical queries
– Faster to get the data back
• MDBs are not a replacement for relational databases– MDBs simply do analytics better!
Measures, Dimensions & Cubes
p. 18
18
• To sum up:– MDBs consist of cubes
– Cubes are made up of measures and dimensions
– Measures are the numbers
– Dimensions are the context for the numbers
– Chapters 4 and 5 will continue this discussion
– Pierluigi Collina is a man of many faces
Measures, Dimensions & Cubes
p. 19
19
• What Is Data Mining?
Next up