Answer: alter row

## Explanation In Azure Data Factory data flows, the **Alter Row transformation** is specifically designed to handle row-level operations including insert, update, upsert, and delete operations when writing to sinks. ### Why Alter Row Transformation is Correct: - The Alter Row transformation allows you to define conditional logic that determines how each row should be processed in the sink - For upsert operations, you can set conditions that specify when a row should be inserted (if it doesn't exist) or updated (if it exists) - This transformation directly supports the upsert pattern required for the scenario - It provides a declarative way to define row-level policies without complex custom coding ### Why Other Options Are Not Suitable: - **Join**: While joins can be used to compare data between sources, they don't inherently define upsert logic for sink operations - **Surrogate Key**: This is a data modeling concept, not a transformation in Azure Data Factory data flows - **Select**: The Select transformation is used for column selection and renaming, not for defining row-level operations on sinks ### Best Practice Consideration: When configuring upsert operations with the Alter Row transformation, you typically define conditions based on key columns that determine whether a record should be inserted (new record) or updated (existing record). This aligns with Azure Data Factory's data flow capabilities for handling slowly changing dimensions and incremental data loading patterns.

Author: LeetQuiz Editorial Team