Extract, transform and load describes transforming data before it reaches the analytical target.
Illustrative scenario
Validate and normalize a source file before loading a reporting table.
Explain your reasoning
Apply the answer to this scenario: When would ELT be preferable? State your assumptions, the expected outcome and how you would check it before changing a real system.
Extract, load and transform describes loading data before transformations in the target platform.
Illustrative scenario
Raw data lands in a warehouse and SQL builds curated tables.
Explain your reasoning
Apply the answer to this scenario: How do you protect raw sensitive data? State your assumptions, the expected outcome and how you would check it before changing a real system.
It is a sequence of steps that moves and transforms data with an explicit input, output and operational contract.
Illustrative scenario
Daily order files become a reporting dataset.
Explain your reasoning
Apply the answer to this scenario: What indicates that a run is complete? State your assumptions, the expected outcome and how you would check it before changing a real system.
Batch processes bounded groups; streaming continuously processes arriving events, often with explicit event-time and lateness rules.
Illustrative scenario
Daily reconciliation differs from a live fraud signal.
Explain your reasoning
Apply the answer to this scenario: What latency does the business actually need? State your assumptions, the expected outcome and how you would check it before changing a real system.
It stores data in flexible forms, commonly in object storage. Governance and table conventions determine its usefulness.
Illustrative scenario
Raw JSON files are retained for reprocessing.
Explain your reasoning
Apply the answer to this scenario: How do you prevent an unsearchable collection of files? State your assumptions, the expected outcome and how you would check it before changing a real system.
Common follow-up
How do you prevent an unsearchable collection of files?
It organizes data for analytical querying with managed storage, compute and governance features depending on the product.
Illustrative scenario
Analysts query curated sales tables.
Explain your reasoning
Apply the answer to this scenario: How does this differ from an operational database? State your assumptions, the expected outcome and how you would check it before changing a real system.
Common follow-up
How does this differ from an operational database?
It identifies the business entity independently of a particular ingestion run.
Illustrative scenario
An order identifier connects updates to the same order.
Explain your reasoning
Apply the answer to this scenario: Can the source reuse that identifier? State your assumptions, the expected outcome and how you would check it before changing a real system.
Repeating an operation has the same intended effect as applying it once.
Illustrative scenario
Reprocessing a file does not duplicate sales.
Explain your reasoning
Apply the answer to this scenario: How do external side effects change the design? State your assumptions, the expected outcome and how you would check it before changing a real system.
It records a progress boundary used to select later work. Its meaning and update timing must be explicit.
Illustrative scenario
Load changes beyond a source sequence.
Explain your reasoning
Apply the answer to this scenario: How will late data behind that boundary be handled? State your assumptions, the expected outcome and how you would check it before changing a real system.
Common follow-up
How will late data behind that boundary be handled?
Unexpected types, missing fields or changed meanings can silently corrupt downstream results.
Illustrative scenario
A numeric amount arrives as localized text.
Explain your reasoning
Apply the answer to this scenario: Which changes can be accepted automatically? State your assumptions, the expected outcome and how you would check it before changing a real system.
It records measurements or events at a defined grain.
Illustrative scenario
One row represents one order line.
Explain your reasoning
Apply the answer to this scenario: Why must grain be stated before joining dimensions? State your assumptions, the expected outcome and how you would check it before changing a real system.
Common follow-up
Why must grain be stated before joining dimensions?
It describes entities used to filter or group facts.
Illustrative scenario
Products have category and brand attributes.
Explain your reasoning
Apply the answer to this scenario: How should historical attribute changes be represented? State your assumptions, the expected outcome and how you would check it before changing a real system.
Common follow-up
How should historical attribute changes be represented?
It can reduce scanned data when queries filter compatible partition keys and the engine can prune partitions.
Illustrative scenario
A report selects one month of events.
Explain your reasoning
Apply the answer to this scenario: How do you confirm actual bytes scanned? State your assumptions, the expected outcome and how you would check it before changing a real system.
It is Google Cloud's analytical data platform for querying large datasets with GoogleSQL and other supported capabilities.
Illustrative scenario
A team analyzes curated transaction data.
Explain your reasoning
Apply the answer to this scenario: How do storage layout and query shape affect cost? State your assumptions, the expected outcome and how you would check it before changing a real system.
Common follow-up
How do storage layout and query shape affect cost?
It coordinates dependencies, schedules and operational state; it does not by itself make tasks correct or idempotent.
Illustrative scenario
An aggregate runs only after its inputs complete.
Explain your reasoning
Apply the answer to this scenario: What happens after a partial upstream failure? State your assumptions, the expected outcome and how you would check it before changing a real system.
They can produce values lazily, avoiding a full in-memory list when processing is naturally incremental.
Illustrative scenario
Read and validate a large text file line by line.
Explain your reasoning
Apply the answer to this scenario: What still needs buffering? State your assumptions, the expected outcome and how you would check it before changing a real system.
A technically successful job can still produce missing, duplicated or invalid business data.
Illustrative scenario
A load completes with half the expected orders.
Explain your reasoning
Apply the answer to this scenario: Which checks detect completeness and correctness? State your assumptions, the expected outcome and how you would check it before changing a real system.
It compares meaningful source and target measures within an agreed scope.
Illustrative scenario
Compare order counts and total amounts for one business day.
Explain your reasoning
Apply the answer to this scenario: How do timezone boundaries affect the comparison? State your assumptions, the expected outcome and how you would check it before changing a real system.
It holds rejected records and error context for diagnosis and controlled correction.
Illustrative scenario
Malformed events are isolated from valid records.
Explain your reasoning
Apply the answer to this scenario: How are repaired records replayed safely? State your assumptions, the expected outcome and how you would check it before changing a real system.
It describes where data came from and how it was transformed.
Illustrative scenario
A report value traces back to a source field and transformation.
Explain your reasoning
Apply the answer to this scenario: How does lineage help assess a schema change? State your assumptions, the expected outcome and how you would check it before changing a real system.