DP-203 Certification Practice Question #151

Question

You are planning a streaming data solution that will use Azure Databricks. The solution will stream sales transaction data from an online store. The solution has the following specifications:
The output data will contain items purchased, quantity, line total sales amount, and line total tax amount.

✑ Line total sales amount and line total tax amount will be aggregated in Databricks.
✑ Sales transactions will never be updated. Instead, new rows will be added to adjust a sale.
You need to recommend an output mode for the dataset that will be processed by using Structured Streaming. The solution must minimize duplicate data.
What should you recommend?

Accepted Answer

The source stream never updates existing sales records; corrections arrive as additional rows. That makes the ingest pattern append-only, and the community majority aligns with using append mode. Microsoft Learn explains that update mode emits rows that changed during a trigger and complete mode emits the full state table, both of which create more churn than necessary for this scenario. Because the question explicitly says adjustments are represented as new rows and asks to minimize duplicates, append is the best fit, though confidence is slightly reduced because some commenters reasonably focus on the downstream aggregate behavior.

More DP-203 practice questions