Certified Specialist Business Intelligence (CSBI)
Practice Exam - 2026/2027 Industry-Aligned
Questions & Verified Answers | HFMA Business
Analytics
Domain 1: Data Architecture & Modeling (Q1-20)
Q1: A retail company needs its BI system to track changes to product categories over
time for accurate historical reporting. Which data-modeling technique is MOST
appropriate?
A. Use a transactional (3NF) schema
B. Implement a Type 1 Slowly Changing Dimension (SCD)
C. Implement a Type 2 Slowly Changing Dimension (SCD)
D. Store only the current state in a flat file
Verified Answer: C
Rationale: Kimball methodology: Type 2 SCD inserts new rows with surrogate keys to
preserve full history, enabling point-in-time analytics. Type 1 (B) overwrites history; 3NF
(A) is OLTP-oriented; flat file (D) lacks structure and referential integrity.
Q2: A Type 3 SCD is BEST suited for:
A. Storing unlimited historical versions
,B. Overwriting history without audit trail
C. Keeping only the current and one previous value in the same row
D. Capturing slowly changing facts in a fact table
Verified Answer: C
Rationale: Type 3 adds a “previous value” column; only two versions are kept. It is not
for unlimited history (A = Type 2) nor for facts (D).
Q3 (SATA): Which characteristics distinguish a Data Lakehouse from a traditional Data
Warehouse? (Select All That Apply)
A. Schema-on-read capability
B. Decoupled compute & storage
C. ACID-compliant transactional layer
D. Mandatory star-schema modeling
E. Support for both structured and unstructured data
Verified Answers: A, B, C, E
Rationale: Lakehouse (e.g., Databricks Delta Lake) adds ACID (C), schema-on-read (A),
cloud-native separation (B), and multi-format support (E). Star schema (D) remains
optional, not mandatory.
Q4: In a star-schema sales data mart, “Ship Date” is added to the fact table. The MOST
appropriate handling is:
,A. Add a second date dimension role-playing key named Ship_Date_Key
B. Create a new fact table only for shipping metrics
C. Add Ship_Date as a degenerate dimension
D. Store Ship_Date as a VARCHAR comment field
Verified Answer: A
Rationale: Role-playing dimension (same date dimension used twice with different
aliases) keeps model compact and performant. Degenerate (C) is for operational
control numbers, not dates.
Q5: A surrogate key is used in dimensional modeling PRIMARILY to:
A. Speed up OLTP inserts
B. Provide a stable, single-field identifier independent of source-system changes
C. Encrypt sensitive business keys
D. Enforce referential integrity in 3NF schemas
Verified Answer: B
Rationale: Surrogate keys (simple integer sequences) insulate the warehouse from
source-system key changes and support SCD types. A is OLTP-centric; C is encryption; D
is relational theory.
Q6 (SATA): Which situations favor Data Vault 2.0 over Kimball star-schema for
enterprise warehousing? (Select All That Apply)
, A. Need for full auditability and historical traceability of every source change
B. High-frequency agile delivery of new data sources
C. Mandatory star-schema for all user-facing tables
D. Multi-country regulatory compliance requiring delete-proof storage
E. Small, single-source, static data mart
Verified Answers: A, B, D
Rationale: Data Vault’s hub-satellite-link design audits every change (A), allows parallel
agile loads (B), and immutable storage (D). C is independent of modeling choice; E is
over-kill.
Q7: A data mesh advocates “domain ownership”. Practically, this means:
A. Central IT team models all enterprise data
B. Each business domain builds and serves its analytical data as a product
C. A single cloud data lake hosts raw files only
D. Governance is optional to speed delivery
Verified Answer: B
Rationale: Data-mesh principle: domains (finance, logistics, HR) own their data products
(pipelines, models, SLA, documentation); central IT provides federated governance
platform.
Practice Exam - 2026/2027 Industry-Aligned
Questions & Verified Answers | HFMA Business
Analytics
Domain 1: Data Architecture & Modeling (Q1-20)
Q1: A retail company needs its BI system to track changes to product categories over
time for accurate historical reporting. Which data-modeling technique is MOST
appropriate?
A. Use a transactional (3NF) schema
B. Implement a Type 1 Slowly Changing Dimension (SCD)
C. Implement a Type 2 Slowly Changing Dimension (SCD)
D. Store only the current state in a flat file
Verified Answer: C
Rationale: Kimball methodology: Type 2 SCD inserts new rows with surrogate keys to
preserve full history, enabling point-in-time analytics. Type 1 (B) overwrites history; 3NF
(A) is OLTP-oriented; flat file (D) lacks structure and referential integrity.
Q2: A Type 3 SCD is BEST suited for:
A. Storing unlimited historical versions
,B. Overwriting history without audit trail
C. Keeping only the current and one previous value in the same row
D. Capturing slowly changing facts in a fact table
Verified Answer: C
Rationale: Type 3 adds a “previous value” column; only two versions are kept. It is not
for unlimited history (A = Type 2) nor for facts (D).
Q3 (SATA): Which characteristics distinguish a Data Lakehouse from a traditional Data
Warehouse? (Select All That Apply)
A. Schema-on-read capability
B. Decoupled compute & storage
C. ACID-compliant transactional layer
D. Mandatory star-schema modeling
E. Support for both structured and unstructured data
Verified Answers: A, B, C, E
Rationale: Lakehouse (e.g., Databricks Delta Lake) adds ACID (C), schema-on-read (A),
cloud-native separation (B), and multi-format support (E). Star schema (D) remains
optional, not mandatory.
Q4: In a star-schema sales data mart, “Ship Date” is added to the fact table. The MOST
appropriate handling is:
,A. Add a second date dimension role-playing key named Ship_Date_Key
B. Create a new fact table only for shipping metrics
C. Add Ship_Date as a degenerate dimension
D. Store Ship_Date as a VARCHAR comment field
Verified Answer: A
Rationale: Role-playing dimension (same date dimension used twice with different
aliases) keeps model compact and performant. Degenerate (C) is for operational
control numbers, not dates.
Q5: A surrogate key is used in dimensional modeling PRIMARILY to:
A. Speed up OLTP inserts
B. Provide a stable, single-field identifier independent of source-system changes
C. Encrypt sensitive business keys
D. Enforce referential integrity in 3NF schemas
Verified Answer: B
Rationale: Surrogate keys (simple integer sequences) insulate the warehouse from
source-system key changes and support SCD types. A is OLTP-centric; C is encryption; D
is relational theory.
Q6 (SATA): Which situations favor Data Vault 2.0 over Kimball star-schema for
enterprise warehousing? (Select All That Apply)
, A. Need for full auditability and historical traceability of every source change
B. High-frequency agile delivery of new data sources
C. Mandatory star-schema for all user-facing tables
D. Multi-country regulatory compliance requiring delete-proof storage
E. Small, single-source, static data mart
Verified Answers: A, B, D
Rationale: Data Vault’s hub-satellite-link design audits every change (A), allows parallel
agile loads (B), and immutable storage (D). C is independent of modeling choice; E is
over-kill.
Q7: A data mesh advocates “domain ownership”. Practically, this means:
A. Central IT team models all enterprise data
B. Each business domain builds and serves its analytical data as a product
C. A single cloud data lake hosts raw files only
D. Governance is optional to speed delivery
Verified Answer: B
Rationale: Data-mesh principle: domains (finance, logistics, HR) own their data products
(pipelines, models, SLA, documentation); central IT provides federated governance
platform.
