Question 140
A financial ETL library transforms standardized customer and order datasets. An auditor must prove the outcome of Run an isolated integration test with controlled input and output tables for this requirement: A pipeline reads real managed tables and the team needs to validate end-to-end behavior. Which test result is best?
Explanation
Correct Answer
The acceptance criterion for “A pipeline reads real managed tables and the team needs to validate end-to-end behavior” is The test verifies actual table reads, writes and expected persisted results; this is the expected evidence after Run an isolated integration test with controlled input and output tables.
Incorrect Answers
This signal belongs to the adjacent case “A table schema must match exact expected field types and nullability,” whose remedy is Use `assertSchemaEqual` against the expected StructType. The observed result for the present requirement is The test verifies actual table reads, writes and expected persisted results.
Tests catch the invalid output for the negative fixtures pertains to a different control (Write focused parameterized unit cases covering null and duplicate inputs) and a different need (A business rule contains edge cases for nulls and duplicates). It cannot establish that Run an isolated integration test with controlled input and output tables met the stated goal.
The unit test detects differences in output records pertains to a different control (Use PySpark `assertDataFrameEqual` with explicit expected rows) and a different need (An expected DataFrame result should be compared to the actual full record content). It cannot establish that Run an isolated integration test with controlled input and output tables met the stated goal.