Inmon advocated building the data warehouse piece by piece as needed
True or False
We need to "reform" dimensions so they can be used by >1 Fact table
True or False
A project plan for an enterprise data warehouse may look like this?
What would be a typical ordering of steps for a dimensional approach to DW development?
Data mart, business requirements, conformance Matrix
Business requirements, conformance Matrix, Data mart
Business requirements, enterprise data warehouse, data mart
Business requirements, Data mart, conformance Matrix
The artifact shown here is a conformance matrix?
True or False
What is the flaw in this artifact
X needed for Professor-Tuition
X not need for Professor-Graduation
Columns for other dimensions are missing
There are no flaws
These two dimensions are not conformed
True or False
Inconsistencies like these could be reconciled automatically
True or False?
The solution to keeping our dimensions conformed is to use Master Data
True or False
In a conformed rollup, one dimension is more granular than the other
In this diagram, Dim1 is the conformed rollup, and Dim2 is the base dimension
True or False
Stakeholders should sign off before developing a data mart without first planning conformed DIMs
True or False
This DIM can hold events like Justin Turner being sacked by the browns
Which is true about how the change in flour was handled in this ProductDIM example?
Type 3 SCD was Used
Type 2 SCD was Used
Type 1 SCD was Used
None of the Above
It is up to IT Professionals to decide whether SCD 1, 2, or 3 should be used
True or False?
Deciding how we want to keep track of history in a data warehouse called "Retrospection"
Which is true about how the change in flour was handled in this ProductDIM example?
Type 3 SCD was Used
Type 2 SCD was Used
Type 1 SCD was Used
Not correct b/c violates 1st normal form
Type 3 SCD is useful for viewing old data through a new lens or vice versa
True or False
If we do not have a natural key, we cannot implement SCD Type 2
True or False
This example illustrates that
SCD 1 generates a lot of new records
SCD 2 generates a lot of new records
SCD 3 generates a lot of new records
Hal Smith should not have been married
There is just one instantiation of the entity in this example
True or False
If we implement a Type 2 SCD for Manager, what will happen?
We will overwrite CustomerAddress
We will change the ManagerKEY
We will add a column for prior manager
We will issue numerous new records for our customers
At any given point in time, some records in a data warehouse will be "out of date"
True or False
Which is true about how the change in flour was handled in this ProductDIM example?
Type 3 SCD was Used
Type 2 SCD was Used
Type 1 SCD was Used
Not correct b/c violates 1st normal form
How can we, as IT people, better ensure high quality data?
All the other choices
Set up communications between data creators and data users
Persuade management to spend $ to fix data at source
Give responsibility for data cleanliness to business owners
Manual entry of website data remains the biggest threat to data quality
True or False
Which ETL principle ensures that the same types of records are always treated the same way?
1st Normal Form
Reliability
Availability
Manageability
Data cubes have a lot of data redundancy to save space
True or False
Which ETL principle ensures that the Data Warehouse can evolve gracefully
Type 2 SCD
Reliability
Availability
Manageability
The function RandInt(n) would be useful for generating the necessary surrogate keys
True or False
When loading a star, we should load the Fact Table first?
True or False
Which ETL principle ensures that Service Level Agreement are met?
3rd Normal Form
Reliability
Availability
Manageability
Inspection of log files is one way to determine when SCD 2 change need to be made
True or False
What is the grain of this FACT table?
Date
Date and Product
Store and Product
Date, Product, and Store
In which of our problem sets have we created Transaction Fact Tables?
PS2 & 3 - Shoplifting Incidents
PS7 - Sales and Targets
PS$ - Emergency Room Check In
All the other choices
What type of FACT table is this?
Event FACTLESS
Transaction FACT
Coverage FACTLESS
Periodic Snapshot FACT
This is a Periodic Snapshot Table
True or False
An accumulating snapshot FACT table tracks what has happened over a period of time
True or False
Select AS Many of the following that are correct
Accumulating snapshot tables use "FAT" data
Periodic snapshot tables use "Thin" data
Transaction FACT tables use "THIN" data
Accumulating snapshot tables use "THIN" data
Accumulating Snapshot FACT tables typically are snowflaked
True or False
Accumulating Snapshot Fact Tables typically use Dimensional Role Play
True or False
A periodic snapshot FACT table captures
One row per transaction
One row for the entire lifetime of an event
One row per time period
None of the other choices
A Prof. that needs to understand student video watching at various points in the semester
Should use a Periodic Snapshot FACT table
Should use a transaction FACT table
Should use an Accumulating Snapshot FACT table
Should get out of the teaching business
Which is true about how the change in flour was handled in this example?
Type 3 SCD was used
Type 2 SCD was used
Type 1 SCD was used
There cannot be columns for Old and New without violating 3rd normal form
A harm that may come from the way firms process PII is that customers may be "outed" as something they either are not, or do not want revealed
True or False
A Periodic Snapshot FACT Table stores the state of a business at a regular time interval
True or False
Accumulating snapshot tables use "FAT" data
True or False
Without a natural (business, persistent) key in a Dimension, we cannot implement SCD Type 2
True or False
There is just one instantiation of the entity in this example
True or False
Data Profiling would be a way to find out that regions did not conform in Problem Set 6
True or False
Stakeholders should make sure that conforming DIMs have been planned before they sign off on (approve) a data mart project
True or False
These two dimensions are not conformed
True or False
You may recall the star schema we used in Problem Set 7. What kind of FACT table is at the center of this STAR?
Periodic Snapshot FACT Table
Conformed Snapshot FACT Table
Accumulating Snapshot FACT Table
Transaction FACT Table
Assume that there can be more than one promotion active for a product on any given day at any given store. One possible way of designing the FACT table would be as follows: Is this Correct?
Yes, because we got rid of the KeyNo as the single PK
No, because PromoDIM should be a causal DIM
No, because it should use dimensional role playing for Promotions
Yes, because it ensures the right granularity
Second Portion Question 1
Over the past few years the owners of this jewelry business received communications from some of their customers who have told them that their product selection does not cater well to people who identify in a non-traditional manner (i.e. non-binary, trans, etc.). Having little personal experience with this, the owners want to make sure they can study the impact it may have on sales, especially for people who first identify one way, but then end up identifying another. What method would you advise the owners to use to capture the needed information?
SCD 3 on the CustomerDIM
SCD 2 on the CustomerDIM
SCD 1 on the CustomerDIM
Dimensional Role Play with the Customer DIM
Second Portion Question 2
This is the DIM they came up with
It fully implements the solution you chose in the prior question
True or False
Second Portion Question 3
Assume that the owners do need an SCD 2 solution (I am not saying this is correct). If that is the case, what do they need to change?
a) Add a Natural Key
b) Add a Surrogate Key
c) Add a Date Stamp
d) Add a Current Flag
e) A, C and D
f) B, C, and D
Second Portion Question 4
Assume that the owners have implemented a Type 2 SCD solution for the CustDIM. What problems might they encounter?
They will have to issue new surrogate keys every time a customer buys a different product
Multiple versions of the same Cust name could cause him to produce a Cartesian Product when joining the Dimension and the Fact table
None of the choices here will be a problem
They will have to use SELECT COUNT (DISTINCT column-name) when using SQL to analyze how many Customers they have
Second Portion Question 5
Another reason the owners want to keep track of this information is because of the possibility that lawsuits might be brought regarding certain products that they sell. Being an enterprise that taps into a wide variety of artists, they sell collections like the "Bad Boy Collection" that might provoke sufficient offense as to lead to a case against them. Or there may be defects with earrings or other pieces that could cause physical injury. They are headquartered in Tennessee, and are aware that the Tennessee courts adopted a star schema system based on Microsoft SQL Server 2005. This is the relevant star they can connect to to find out if any cases have been filled regarding the products they carry:
What kind of FACT table is present in this star?
Accumulating Snapshot FACT Table
Product Bridge FACT Table
Periodic Snapshot FACT Table
Event Factless Fact Table
Second Portion Question 6
Compare the Product dimensions in each of the two stars (the original Sales star and the new courts star). Pick as many of the following statements as you think are correct:
A. The two dimensions are not conformed because the fields are not identical.
B. The two dimensions are not conformed because the presence of a Bridge table makes the relationship between the two stars uncertain.
C. The two dimensions are conformed using a conformed roll-up.
D. The two dimensions are not conformed because the granularity in the Disposition FACT table is not correct.
E. The two dimensions are conformed because they each have a Product Category.
F. The two dimensions are not conformed because the contents of the fields are not identical.
G. The two dimensions are conformed because they are identical.
Second Portion Question 7
Assume there is a conformance problem with the ProductDIM.
TF: This vest way to address this problem is by using a Conformance Matrix
True
False
Second Portion Question 8
The owners of the Jewelry business now want to check to see which, if any, of their products may have been involved in Tennessee court cases. Assuming that any conformance issues with the ProductDIM have been resolved (if there are any), on what dimension(s) should the owners drill across?
ProductDIM-SalesFACT Star to ProductDIM DispositionFACT Star
Both ProductDIM-SalesFACT Star to ProductDIM DispositionFACT Star and DateDIM-SalesFACT Star to DateDIM DispositionFACT Star
The bridge table in the DispositionFACT Star makes it impossible to drill across
Sales FACT to DispositionFACT
DateDIM-SalesFACT Star to DateDIM DispositionFACT Star
Second Portion Question 9
The Tennessee Court system has noticed that some of the records in the Star Schema for disposition of cases are missing data for some fields. This is hindering its ability to do important analysis. What would be the best single thing to do (if you can only do one of these) to ensure the data is complete and accurate?
Make sure that the business side has responsibility for data quality
Hire extra court stenographers to take down all the information
Set up IT systems that will reject any data that does not fit the Master Data files
Make the lawyers on each side responsible for entering their own information through a portal