SF-PD2 Performance Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Retrieve all records and aggregate them in an Apex loop.

  Why it's wrong here

  Pulling 10 million rows into Apex memory will immediately hit the heap size limit and cause the transaction to fail. Aggregating in memory is extremely inefficient compared to letting the database engine handle the heavy lifting, and it is a major performance anti-pattern in large-scale data processing.

• ✓

  Use SOQL with GROUP BY to perform aggregation in the database.

  Why this is correct

  Using the GROUP BY clause allows the database engine to perform the aggregation, which is highly optimized. This approach avoids transferring large volumes of data to the Apex heap, effectively preventing memory and CPU limit errors while providing the results in a much faster, more efficient manner.

• ✗

  Process records in small batches using a flow.

  Why it's wrong here

  Flows are not suitable for processing 10 million rows; the overhead of the flow engine would make this process prohibitively slow. Batch Apex is the correct programmatic tool, but even then, performing the aggregation via SOQL is significantly faster than doing so within the Batch Apex logic.

• ✗

  Use a custom index on all fields involved in the query.

  Why it's wrong here

  While indexing is important for filtering, it does not perform the actual aggregation. Aggregating data still requires processing the records. A custom index helps locate records quickly, but the logic must still use the GROUP BY clause to achieve efficient, database-level aggregation of the result set.