ETL Testing Basic 4

Question 1

Change Data Capture (CDC) in ETL primarily captures:

Accepted Answer

Only the rows that were inserted, updated, or deleted since the last extraction

Answer

CDC identifies and extracts only changed records, reducing extraction volume and enabling near-real-time data warehouse updates.

Question 2

Data reconciliation testing in ETL is best described as:

Accepted Answer

Comparing source and target record counts and key metrics to confirm no data was lost or duplicated during the ETL process

Answer

Reconciliation compares totals, counts, and checksums between source and target to certify that every expected record landed correctly.

Question 3

Which approach best prevents duplicate records when running incremental ETL loads?

Accepted Answer

Using a high-watermark or last-run timestamp combined with an UPSERT (merge) operation

Answer

A high-watermark limits extraction to new/changed rows, and an UPSERT ensures existing target rows are updated rather than re-inserted.

Question 4

In ETL testing, a 'boundary value' test would check:

Accepted Answer

Data at the edges of allowed ranges, such as the minimum, maximum, and just-outside-maximum values for a numeric field

Answer

Boundary value testing targets the extreme valid and just-invalid values of a field to confirm that validation logic accepts or rejects them correctly.

Question 5

What is 'metadata testing' in the context of ETL?

Accepted Answer

Validating that column names, data types, lengths, and constraints in the target match the agreed-upon design specifications

Answer

Metadata testing checks structural correctness — ensuring the target schema (column names, types, nullable flags, etc.) matches the mapping specification.

ETL Testing - extract, transform, load Practice Test

ETL Testing - extract, transform, load Practice Test

ETL Testing Basic 4