Data Warehouse and Data Marts — Learn free

Self-check

Questions 1–10 of 10

1. Why did the idea of a data warehouse emerge?
Question 1A. Because OLTP systems could not store structured dataB. Because transactional systems were never designed for heavy reporting or analytical workloadsC. Because businesses wanted to delete historical recordsD. Because relational databases were being replaced by NoSQLE. Because cloud storage became free
2. What is the distinction between a data warehouse and dimensional modelling?
Question 2A. They are two names for the same thingB. A data warehouse is a technique; dimensional modelling is a repositoryC. A data warehouse is the central repository consolidating data from many sources, while dimensional modelling is one possible technique for structuring itD. Dimensional modelling is used only in OLTP, never in a warehouseE. A data warehouse must always use dimensional modelling
3. Which data warehouse modelling approach, popularised by Bill Inmon, emphasises integration and normalisation to minimise redundancy?
Question 3A. Star schemaB. Third Normal Form (3NF)C. Data VaultD. OLAP cubeE. Galaxy schema
4. In dimensional modelling, what does a fact table keep?
Question 4A. The data points by which analysis is sliced, such as time or regionB. The transaction's lowest level of data, numeric or non-numeric, such as sales numbers or clicksC. The metadata describing each source systemD. Hubs, links and satellitesE. Pre-aggregated multidimensional totals
5. What is a conformed dimension?
Question 5A. A dimension that has been normalised into related tablesB. A dimension that can connect to multiple factsC. A dimension stored in a data mart onlyD. A dimension that never changesE. A dimension holding only numeric values
6. Which modelling approach structures data into hubs, links and satellites to support historical tracking and auditability?
Question 6A. Snowflake schemaB. Anchor modellingC. Data VaultD. Hook modelE. Third Normal Form
7. What is the main difference between a data warehouse and a data mart?
Question 7A. A warehouse uses dimensional modelling while a mart uses 3NFB. A warehouse allows deletions while a mart does notC. A warehouse consolidates data at enterprise level, while a mart is scoped to a department or function such as Finance or SalesD. A warehouse is virtual while a mart is physicalE. A warehouse stores raw data while a mart stores only cubes
8. How did Bill Inmon define a data warehouse in 1990?
Question 8A. A copy of transaction data specifically structured for query and analysisB. A subject-oriented, integrated, time-variant and non-volatile collection of data in support of management's decision-making processC. A centralised repository of raw data in any formatD. A set of hubs, links and satellitesE. A virtual layer over distributed sources
9. How do the approaches of Bill Inmon and Ralph Kimball differ in building order?
Question 9A. Inmon builds marts first (bottom-up); Kimball builds the enterprise warehouse first (top-down)B. Inmon builds the enterprise warehouse first (top-down); Kimball starts with business-driven data marts (bottom-up)C. Both build marts first but differ on schema typeD. Both build the warehouse first but Kimball uses 3NFE. Inmon rejects data marts entirely
10. Why does the author of I Am Datapedia! prefer building the enterprise warehouse before the data marts?
Question 10A. Because marts cannot use dimensional modellingB. Because regulators require a warehouse firstC. Because if marts are built first, any change at warehouse level forces modifications in marts and upstream pipelines, adding maintenance overheadD. Because warehouses are cheaper to store than martsE. Because Kimball recommended it