Data Warehouse - Mini Quiz

What is the main purpose of a data warehouse?

• To process high volumes of day-to-day transactions quickly (OLTP)

• To store and analyse large volumes of historical data for decision-making (BI)

• To store only unstructured social media posts

• To replace all operational databases

“Subject-oriented” means a data warehouse is organised around:

• Individual transactions (e.g., one sale at a time)

• Business subjects/themes (e.g., sales, customers, finance)

• File types (CSV, JSON, log files)

• Random partitions for speed

A key feature of “integrated data” is that:

• Data is kept separate in each original system

• Data from many sources is combined into one consistent repository

• Only one table is used to store everything

• Data is deleted regularly to save space

Which term best matches “stores historical snapshots for trend analysis”?

• Non-volatile

• Time-variant

• Subject-oriented

• Append-only

“Non-volatile” data in a warehouse means:

• Data is constantly updated like an operational database

• Data is deleted when it becomes old

• Once loaded, data is rarely changed or deleted

• Data is stored only in RAM

“Append-only” means:

• Existing records are edited in place

• New data is added as new rows; old records are not overwritten

• Data is stored only in external files

• Data cannot be queried

Which pair best explains why warehouses support reliable historical analysis?

• Append-only + Non-volatile

• Subject-oriented + Volatile

• Integrated + Randomised

• Transactional + Real-time

“Optimised for query performance” means a data warehouse is designed to:

• Prioritise fast INSERT/UPDATE operations

• Run complex analytical queries efficiently across huge datasets

• Prevent the use of indexes

• Avoid aggregation functions

Which statement best contrasts OLTP and a data warehouse?

• OLTP stores only historical data; warehouse stores only current data

• OLTP is read-heavy; warehouse is write-heavy

• OLTP focuses on fast transactions; warehouse focuses on analytics and trends

• They are identical but use different names

A warehouse is “integrated” partly because data is cleaned/transformed to a consistent format (often via ETL). Which is the best description of ETL?

• Encrypt, Transfer, Log

• Extract, Transform, Load

• Evaluate, Test, Launch

• Encode, Translate, Link