Quiz 5 Flashcards
1
Q
multidimensional data
A
2
Q
dimension
A
attribute of decision-making
3
Q
fact
A
outcome measure
4
Q
dimension table key
A
- Uniquely identifies a record
- Should not be same as operational database key
- Avoid built-in meanings into the key used in the data warehouse
5
Q
dimension table records
A
- Few records in table
6
Q
dimension table attributes
A
- Large number of attributes
- Attributes capture specific aspects of decisions made
- Mostly non-numeric, categorical data types
7
Q
operational vs. DW primary keys
A
-
Operational Keys:
- Meaningful
- Unique at that time and while it is present in the database
- Key can change
- New column is defined as key
- Data type changed
-
Surrogate Key (DW):
- Completely independent of operational
- Completely meaningless
- Operational key kept in data warehouse but as a regular attribute, not as a primary key
8
Q
fact table key
A
- concatenated key of dimension keys
- only two types of columns in fact table
- concatenated keys
- measures
9
Q
fact table granularity
A
- Determined by Granularity of Dimension Tables
- Example:
- Each record represents sales by (Dealer, Date, Customer, Product)
10
Q
fact table measures
A
- typically numeric attributes
11
Q
fact table sparseness
A
- Small percentage of combinations of dimensions are recorded in the fact table
12
Q
fact table size
A
- Compared to dimension table, this table is very large
- >100x smallest dimension table
13
Q
(fully) additive measure
A
- Measure that is summed during roll up
- Stored in fact table
- Example:
- Total Sales = ∑Sales<em>i</em>
14
Q
semi-additive measure
A
- Calculated
- All of its components are additive, but itself is not
- Components stored but result calculated on the spot
- Example:
- Average Sales = Total Sales / Count of sales
15
Q
non-additive measure
A
- Calculation must always be performed starting at finest granularity
- If any component is non-additive, then the measure is non-additive
- Example:
- Total number of unique customers
- No way to determine from roll-up
16
Q
factless fact table
A
- Has no measures
- Each record identifies occurrence of an event
- Example:
- Test-drive of a car
- Which customer, When (Time), Where (Dealer), What Car (Product)
-
Why?
- Analyze influence of events on Other Outcomes
- For example:
- Do certain promotions increase test-drives?
- Relationship between test-drives and sales
17
Q
star schema and queries
A
- star schema makes querying simple
18
Q
star schema cost of query
A
- …
19
Q
roll-up
A
- A roll-up involves summarizing the data along a dimension
- The summarization rule might be computing totals along a hierarchy or applying a set of formulas such as “profit = sales - expenses”
20
Q
drill-down/up
A
- Drill Down/Up allows the user to navigate among levels of data
- ranging from the
- most summarized (up) to the
- most detailed (down)

21
Q
slice
A
- Slice is the act of picking a rectangular subset of a cube by choosing a single value for one of its dimensions
- creating a new cube with one fewer dimension
