Is Data Warehouse Transaction Oriented or Query Oriented?

Question

Context of the Problem

Reading on the subject Data Warehouse, I have drawn some distinct conclusions about what the text book says.

Note what the author explains regarding transaction-oriented DW:

I understand that the author claims that the DW are transaction-oriented.

Now see what he says in the following parable:

Right after he claims that the DW are not transaction oriented routine, which got me a little confused, since he’s denying.

My Understandings

1. When reading about the subject I understood that DW are decision support tools and that are consultation-oriented.

2. I understood that by being focused on the consultations, the bank may not respect the well-known normal forms (1FN,2FN,3FN), since they increase the amount of relationships and diminish the chance of duplication.

3. The fact is that the relationship building at the bank affects performance of the consultations because of the various JOIN that are constantly held.

4. DW is usually quite large since it serves as the central repository of all other databases.

5. Transactional database (Conventional) are concerned about the nonduplication register, thus, tend to respect the normal forms. They can be called banks transaction-oriented.

6. Data stored in DW has feature NONVOLATILE, means to say that in the loading step, the data is filtered and cleaned, after that step the data only suffer operations of consultations and exclusions, sem que possam ser alterados, this characteristic represents the nonvolatility.

Doubt

Are my understandings on the subject correct? The community fully or partially agrees with the sentences listed above?

Recalling that the main focus of the question is in the title, that is, I can consider in fact that DW are oriented to consultation?

Source: Database Systems 6th edition, authors: Elmasri, Ramez Navathe, Shamkant B. Year: 2011 Cap 29, pg 720

Answer 1

I found the first half-short fragment, it is worth remembering another important characteristic of DW that the data practically do not suffer (nor should) changes to not interfere in the result of the analysis and new records are added periodically via lot (ETL).

Transactions are used mostly in write operations where it is important to complete all changes successfully or cancel all and return to previous (valid) status.