CSI Data Mapping & Transformation 2 — Questions and Answers
Question 1: Which transformation pattern converts a single source record into multiple target records?
- One-to-one mapping
- One-to-many splitting (Correct answer)
- Many-to-one aggregation
- Passthrough transformation
Correct answer: One-to-many splitting
One-to-many splitting (or explosion) breaks a single source record into multiple target records, such as expanding a line-item order into individual product records.
Question 2: In XSLT-based data transformation, what is the primary role of a stylesheet?
- Storing XML data persistently
- Defining rules for transforming XML documents into another format (Correct answer)
- Validating XML against a schema
- Encrypting XML payloads in transit
Correct answer: Defining rules for transforming XML documents into another format
An XSLT stylesheet defines templates and transformation rules that convert XML source documents into a target format such as HTML, another XML structure, or plain text.
Question 3: When performing a many-to-one aggregation in data mapping, which operation is most commonly applied?
- Filtering duplicate records
- Combining multiple source records into a single target record using a function like SUM or COUNT (Correct answer)
- Splitting fields by delimiter
- Reformatting date strings
Correct answer: Combining multiple source records into a single target record using a function like SUM or COUNT
Many-to-one aggregation consolidates multiple source records into a single target record, typically using aggregate functions such as SUM, AVG, MAX, or COUNT.
Question 4: A canonical data model in enterprise integration is best described as:
- A physical database schema shared across all applications
- A common, application-neutral data format used as an intermediary for message translation (Correct answer)
- A data dictionary listing all field names in each system
- A compressed binary format for high-throughput messaging
Correct answer: A common, application-neutral data format used as an intermediary for message translation
A canonical data model (CDM) is a vendor-neutral, agreed-upon intermediate format that decouples source and target systems, reducing the number of point-to-point transformations needed.
Question 5: Which data quality issue occurs when a field expected to hold a date contains a string like 'N/A'?
- Schema drift
- Data type mismatch (Correct answer)
- Referential integrity violation
- Duplicate key error
Correct answer: Data type mismatch
A data type mismatch occurs when a value stored in a field does not conform to the expected data type, such as text appearing in a date column.
Question 6: In ETL pipelines, what is the purpose of a surrogate key introduced during the transformation phase?
- To encrypt the natural primary key from source systems
- To provide a system-generated, unique identifier independent of source system keys (Correct answer)
- To concatenate composite keys into a single string
- To replace NULL values in foreign key columns
Correct answer: To provide a system-generated, unique identifier independent of source system keys
Surrogate keys are system-generated integers or GUIDs assigned in the data warehouse to uniquely identify records regardless of source system key formats or changes.
Question 7: Which approach is used to handle schema evolution in data transformation pipelines with minimal disruption?
- Hard-coding field positions in every transform script
- Using flexible schema registries and forward/backward-compatible serialization formats (Correct answer)
- Stopping the pipeline and rewriting all mappings
- Replacing the target system's schema to match the source
Correct answer: Using flexible schema registries and forward/backward-compatible serialization formats
Schema registries (e.g., Confluent Schema Registry) combined with formats like Avro or Protobuf support forward and backward compatibility, allowing schemas to evolve without breaking existing consumers.
CSI Data Mapping & Transformation 2
Which transformation pattern converts a single source record into multiple target records?