Errata

Figure 12-2 shows an RDW in a Data Lakehouse Architecture

Note from the Author or Editor:
Figure 12-2 needs to be fixed: "RDW" should be replaced with "Relational Serving Layer (optional)". "Schema on write" should be replaced with "Schema on read". The two icons in the "Schema on write" box, the document and star, should be removed.

Type 3 SCD is described as follows in the book: Type 3 SCDs create a new record for each change so you can maintain a complete history of the data. This type of SCD is the most complex but also the most flexible. For example, when a customer moves from one US state to another, a Type 3 SCD would store the entire old record and an entire new record, which could include dozens of fields and not just the US state field.

However, it is commonly accepted that Type 3 SCD adds a new field/attribute/column to keep track of historical changes, instead of adding new records as described in the book. A type 3 SCD will add a new column to keep track of the latest change of an attribute only. In contrast, a type 2 SCD will add new records to keep track of the entire history of a dimension.

Note from the Author or Editor:
What it should read:

Type 2
Type 2 SCDs maintain multiple versions of the data: the new data and a record of the old data. Use this type of SCD when you need to track the changes to the data over time and maintain a record of the old data. For example, let’s say that six months ago, a company analyzed its sales data and found New York to be its top-selling state. Now, if some customers have since moved out of New York to New Jersey, rerunning the same report without accounting for these changes would show inaccurately lower sales figures for New York. This would lead to a mistaken perception of historical data showing declining sales in New York, which could influence strategic decisions. So, if a customer moves from one US state to another, the company’s Type 2 SCD would create a new record to store both the old state and the new one.

Type 3
Type 3 SCDs add a new column for each change so you can maintain a limited history of the data. This type of SCD is less complex but also more limited in scope. For example, when a customer moves from one US state to another, a Type 3 SCD would add a new column to store the previous state, while the current state is updated in the existing field.