Sample question 2 — Describe considerations for working with non-relational data on Azure
Northwind Traders runs an order-entry system on Azure SQL Database. Every order row stores the buyer's full postal address, so a single corrected street name must be edited on hundreds of rows and some rows retain the old value. The database team wants the address kept once and referenced from each order. Which schema change should the team make?
- A. Archive older orders to Azure Blob Storage and query them later with Synapse.
- B. Normalize the design so addresses live in a customer table referenced by key.
- C. Replicate the orders table to a second Azure region for read-only reporting.
- D. Add a covering index on the address columns so lookups return rows faster.
Answer: B. Option B is right because normalization removes the repeated address values from every order row and keeps one authoritative copy in a customer table that each order references through a key. Correcting a street name then becomes a single update, and no stale duplicates remain. Option A only relocates old rows for later analysis, so live order edits still touch many records. Option C copies the same flawed structure to another region and adds no uniqueness. Option D speeds lookups but leaves the repeated values and update anomalies in place.