Courseiva

CCNA Implement Manage Semantic Models Questions

62 questions · Implement Manage Semantic Models topic · All types, answers revealed

1
MCQeasy

You are creating a semantic model in Microsoft Fabric. You need to ensure that users can only see data for their own region when they view reports. What should you implement?

A.Object-level security (OLS) on the Region column.
B.Row-level security (RLS) with DAX filters.
C.Sensitivity labels applied to the semantic model.
D.Workspace roles with viewer permissions.
AnswerB

Row-level security (RLS) allows you to define DAX filters that restrict data based on the user's identity. By creating roles and applying filters such as [Region] = USERPRINCIPALNAME(), you can ensure users only see data for their region. This is the standard method for implementing data-level security in Power BI and Fabric semantic models.

Why this answer

Row-level security (RLS) with DAX filters is the correct approach to restrict data based on user identity. It allows dynamic filtering so users see only their region's data. OLS, sensitivity labels, and workspace roles do not provide row-level data filtering.

Exam trap

The trap here is confusing object-level security with row-level security, but only RLS can filter rows based on user identity.

2
MCQmedium

You are the analytics engineer for a retail company using Microsoft Fabric. The semantic model contains a large fact table named 'Sales' and a dimension table named 'Product'. Users frequently run queries that filter by product category and aggregate sales amounts. You notice that queries performing these aggregations are slow. You decide to create a user-defined aggregation table to improve performance. Which of the following must you do to ensure the aggregation table is used by the semantic model?

A.Use the 'Manage aggregations' dialog to map each column in the aggregation table to the corresponding column in the detail table and set the aggregation function.
B.Set the aggregation table's storage mode to Import and configure the 'Alternate of' property for each column to map to the corresponding detail table columns.
C.Create a relationship between the aggregation table and the detail table on the key columns, and mark the relationship as active.
D.Add a calculated column to the detail table that computes the aggregated value, and then create a measure that references that column.
AnswerA

The 'Manage aggregations' dialog in Power BI Desktop is specifically designed to define how an aggregation table maps to a detail table. You specify which columns are used for grouping and which measures are aggregated, along with the aggregation function (Sum, Count, etc.). This mapping allows the engine to automatically redirect queries to the aggregation table when appropriate.

Why this answer

To enable automatic query redirection to an aggregation table, you must use the 'Manage aggregations' feature in Power BI Desktop. This dialog allows you to specify the mapping between the aggregation table and the detail table, including which columns are used for grouping and which measures are aggregated, along with the aggregation functions. Once configured, the engine can transparently use the aggregation table for queries that match the defined grain, improving performance without requiring changes to reports.

Exam trap

The trap here is assuming that simply creating a relationship between the aggregation table and the detail table is sufficient for the engine to use the aggregation table, when in fact you must explicitly configure aggregations through the 'Manage aggregations' dialog.

3
MCQmedium

You have a Direct Lake semantic model connected to a Fabric Lakehouse. During a peak usage period, the model exceeds the memory limits of the Fabric capacity. What will be the behavior of the semantic model in this situation?

A.The model will stop responding and return an 'Out of Memory' error to all users.
B.The model will automatically switch to DirectQuery mode.
C.The Fabric capacity will automatically scale up to the next available SKU.
D.The model will compress the existing data further using a higher compression ratio.
AnswerB

This is the 'fallback' behavior of Direct Lake. If the memory required to 'page in' the Parquet files exceeds the allocated capacity, the engine reverts to DirectQuery. In this mode, it queries the data through the SQL endpoint. This transition is automatic and intended to keep reports active, albeit at a lower performance level than pure Direct Lake.

Why this answer

Direct Lake mode has a built-in fallback mechanism to maintain report availability. When the capacity's memory limits are reached, the engine cannot load more Parquet data into memory. To ensure users can still see data, the model transparently switches to DirectQuery mode, sending queries to the SQL Analytics Endpoint of the Lakehouse.

This prevents total failure but may significantly impact query performance.

Exam trap

Candidates incorrectly assume that exceeding memory limits in a Direct Lake model causes the refresh to fail entirely or throw a hard out-of-memory error.

4
MCQhard

You are optimizing a semantic model in Microsoft Fabric that contains a large fact table with billions of rows. Queries that aggregate sales by month and year are slow. You decide to implement a hybrid table approach. What is the primary benefit of using hybrid tables in this scenario?

A.They allow real-time data ingestion without affecting query performance.
B.They enable the use of calculation groups without additional configuration.
C.They store historical data in Import mode and recent data in DirectQuery mode, improving performance for common queries.
D.They automatically create aggregations for all numeric columns.
AnswerC

Hybrid tables partition the table into a historical portion stored in Import mode (in-memory) and a recent portion in DirectQuery mode. This allows fast query performance for historical data while providing near real-time access to recent data. Queries that aggregate over historical periods benefit from the in-memory storage, reducing latency.

Why this answer

Hybrid tables improve performance by storing historical data in Import mode (in-memory) and keeping recent data in DirectQuery mode. This means queries that aggregate over historical periods can be satisfied quickly from memory, while recent data is queried directly from the source. This approach balances performance and data freshness, making it ideal for large fact tables with time-based queries.

Exam trap

The trap here is assuming that hybrid tables provide real-time ingestion or automatic aggregations, when their main purpose is to optimize performance by partitioning data between Import and DirectQuery modes.

5
Multi-Selectmedium

You are using the Tabular Object Model (TOM) and the XMLA endpoint to automate the management of semantic models. You need to update the descriptions of several measures to improve data discovery. Which TWO tools or methods can you use to connect to the XMLA endpoint and modify these properties?

Select 2 answers
A.Tabular Editor 2 or 3
B.SQL Server Management Studio (SSMS)
C.Power BI Service 'Data Hub' interface
D.Fabric Notebooks using the PySpark engine
E.The 'Transform Data' (Power Query) button in the Fabric portal
AnswersA, B

Tabular Editor is a lightweight and powerful tool that connects via the XMLA endpoint. It allows you to quickly edit metadata, such as measure descriptions, calculation groups, and display folders, without having to download the PBIX file. It is the preferred tool for many Fabric engineers for batch updates and metadata management.

Why this answer

Tabular Editor 2 or 3 (A) is correct because it is a dedicated TOM-based client that connects directly to the XMLA endpoint of a semantic model and allows editing of measure properties such as Description, then saving the changes back to the model. SQL Server Management Studio (B) is correct because SSMS can connect to the XMLA endpoint and, via XMLA/TMSL scripts or the tabular model scripting capabilities, modify measure metadata including descriptions. The Power BI Service 'Data Hub' interface (C) is only a discovery/browsing surface and does not provide a way to edit measure properties through XMLA.

Fabric Notebooks using the PySpark engine (D) operate on Spark/Delta workloads and do not natively connect to the XMLA endpoint to alter TOM measure metadata. The 'Transform Data' (Power Query) button (E) opens Power Query for data transformation and does not expose TOM measure description editing over XMLA.

Exam trap

DP-600 often tests which tools can write TOM metadata via XMLA — candidates incorrectly include Power BI Service UI or Fabric notebooks, which cannot modify measure descriptions through the XMLA endpoint.

6
MCQmedium

You are building a composite model in Microsoft Fabric. You have an existing 'Corporate Sales' semantic model published in the service, and you need to combine it with a local Excel file containing 'Target' data. What is a key consideration when using DirectQuery for Power BI semantic models in this scenario?

A.The local Excel file must be uploaded to OneLake before it can be used in a composite model.
B.You cannot use Row-Level Security (RLS) in a composite model that includes DirectQuery.
C.All tables from the 'Corporate Sales' model must be imported into the local model.
D.The 'DirectQuery for Power BI datasets and AS' feature must be enabled in Power BI Desktop.
AnswerD

This feature is what enables the creation of a composite model by allowing a DirectQuery connection to an existing semantic model. Without this enabled, you are limited to a 'Live Connection,' which prevents you from adding any other data sources or tables to the model.

Why this answer

When connecting to a Power BI semantic model to create a composite model, you are not converting a 'Live Connection' to DirectQuery; rather, you are establishing a DirectQuery connection to the semantic model from the start. This allows the model to be extended with additional data sources (like the Excel file). Option D is the correct requirement for enabling this functionality in Power BI Desktop.

Exam trap

Candidates often confuse 'Live Connection' (which is restrictive) with 'DirectQuery for Power BI datasets' (which allows for composite models).

7
MCQeasy

You are building a semantic model and notice that the 'Order' table and the 'Customer' table both have a 'CustomerID' column. You create a relationship between them. By default, what is the cardinality and the most likely direction of the filter in this relationship?

A.Many-to-many, bidirectional
B.One-to-many, single direction
C.One-to-one, bidirectional
D.Many-to-one, single direction
AnswerB

This is the standard and default relationship type. The 'One' side is the Customer table (where each ID is unique), and the 'Many' side is the Order table. Filter context naturally flows from the Customer to the Order table, allowing you to see all orders for a specific customer, which is the most common analytical requirement.

Why this answer

When you create a relationship between an 'Order' table and a 'Customer' table on 'CustomerID', the default cardinality is one-to-many (one customer to many orders) with a single direction filter from the 'Customer' (one side) to the 'Order' (many side). This is the standard default in Power BI and Fabric semantic models.

Exam trap

DP-600 often tests the default relationship settings, and candidates may confuse the direction or cardinality, especially when the tables have a common column name like 'CustomerID'.

How to eliminate wrong answers

Option A is wrong because many-to-many bidirectional is not the default; it requires explicit configuration and is used for specific scenarios like bridge tables. Option C is wrong because one-to-one bidirectional is not typical for a customer-order relationship; it would imply each customer has exactly one order, which is not the case. Option D is wrong because many-to-one single direction is the reverse of the default; the default is one-to-many, not many-to-one, and the filter direction is from the one side to the many side.

8
MCQmedium

You have a semantic model that uses a large fact table. You want to implement a hybrid table approach where the most recent data is available in real-time via DirectQuery, while historical data is stored in Import mode for performance. Which feature must you configure to enable this?

A.Calculation Groups with a 'Real-time' item.
B.Direct Lake fallback settings.
C.Incremental refresh with the 'Get the latest data in real time' option.
D.Dual storage mode on the fact table.
AnswerC

By enabling Incremental Refresh and checking the 'Get the latest data in real time with DirectQuery' box, Power BI creates a hybrid table. The historical partitions are imported into memory, while the partition containing the latest data is queried using DirectQuery, ensuring real-time visibility for new records.

Why this answer

Incremental refresh with the 'Get the latest data in real time' option is the specific Power BI feature that enables hybrid tables. It partitions the table so historical partitions are stored in Import mode (compressed, in-memory) while the most recent partition is set to DirectQuery, allowing real-time querying of the latest data. This is the only native mechanism that produces a hybrid table in a Power BI semantic model.

Exam trap

DP-600 often tests the confusion between hybrid tables (incremental refresh with real-time DirectQuery) and Dual storage mode, which is a dimension-table optimization, not a fact-table hybrid pattern.

How to eliminate wrong answers

Option A is wrong because calculation groups only modify DAX measure behavior at query time; they cannot change how data is stored or queried and have no 'Real-time' item type. Option B is wrong because Direct Lake fallback settings control whether a Direct Lake model falls back to DirectQuery or Import when it cannot read Delta tables directly — they do not create a hybrid Import/DirectQuery table. Option D is wrong because Dual storage mode applies to dimension tables in a model with a DirectQuery fact table; it caches the dimension in Import mode while keeping it queryable in DirectQuery, but it does not split a fact table into real-time and historical portions.

9
MCQmedium

Your organization requires that users only see data from their own department. You decide to use a static Row-Level Security (RLS) role. How should you identify the users for these roles to ensure the security model is scalable?

A.Hardcode individual user email addresses within the DAX filter expression.
B.Assign Microsoft Entra security groups to the roles in the workspace.
C.Create a separate semantic model for every department.
D.Implement object-level security on every table in the model.
AnswerB

Mapping security groups to RLS roles allows you to manage access at the identity provider level. This creates a scalable security framework where administrative overhead is minimized, and access governance is centralized, aligning with enterprise best practices for identity management and data security within the Microsoft Fabric environment.

Why this answer

Assigning Microsoft Entra security groups to RLS roles in the workspace is the most scalable approach because it centralizes user management in Entra ID (formerly Azure AD). Instead of maintaining individual user mappings in the semantic model, you manage group membership in Entra, and any changes automatically propagate to the RLS role. This reduces administrative overhead and ensures consistency across multiple models and reports.

Exam trap

DP-600 often tests the scalability of RLS implementations, and candidates may incorrectly choose hardcoded user lists or separate models instead of leveraging Entra groups.

How to eliminate wrong answers

Option A is wrong because hardcoding individual user email addresses in DAX filter expressions is not scalable; adding or removing users requires editing the DAX and republishing the model, and it is error-prone. Option C is wrong because creating a separate semantic model for every department duplicates data and maintenance, leading to high overhead and potential inconsistencies. Option D is wrong because object-level security (OLS) restricts access to tables or columns, not rows; it does not address the requirement of filtering data by department.

10
MCQmedium

Your organization uses a star schema for its semantic models. You have a requirement to ensure that users can only see the 'Unit Cost' column if they belong to the 'Finance' security group. Users not in this group should still see the rest of the 'Products' table. Which feature should you implement?

A.Row-Level Security (RLS) with a dynamic filter.
B.Object-Level Security (OLS) on the 'Unit Cost' column.
C.Calculation groups with a security measure.
D.A separate perspective for non-finance users.
AnswerB

OLS allows you to secure specific columns or tables. When OLS is applied to the 'Unit Cost' column for a specific role, users in other roles will not see the column in their field list, and any report visuals using that column will return an error for those unauthorized users.

Why this answer

Object-Level Security (OLS) restricts access to specific tables or columns within a semantic model based on user identity or group membership. Since the requirement is to hide only the 'Unit Cost' column from non-Finance users while keeping the rest of the 'Products' table visible, OLS is the correct feature. RLS filters rows, not columns, so it cannot satisfy this requirement.

Exam trap

DP-600 often tests the distinction between RLS (row filtering) and OLS (object/column hiding) — candidates frequently pick RLS because it is more familiar, missing that column-level restrictions require OLS.

How to eliminate wrong answers

Option A is wrong because Row-Level Security filters rows based on DAX predicates and cannot hide an entire column from specific users. Option C is wrong because calculation groups apply time intelligence or measure transformations and have no native security function for hiding columns. Option D is wrong because perspectives only organize visible objects for report authors; they do not enforce security and users can still query hidden objects via DAX or XMLA.

11
MCQmedium

You are managing a semantic model where users need to perform calculations based on a dynamic time period. Which DAX function should you use to create a time-intelligence calculation that respects the current filter context?

A.Use the FILTER function with manual date comparisons.
B.Use the CALCULATE function with time-intelligence helpers.
C.Use the ADDCOLUMNS function to create a date hierarchy.
D.Use the LOOKUPVALUE function to pull values from previous years.
AnswerB

CALCULATE is the engine's primary function for modifying filter context. When paired with time-intelligence helpers, it allows for seamless manipulation of the date range based on the current selection. This is the standard, best-practice approach for implementing year-to-date, previous year, or rolling window calculations within any Power BI semantic model.

Why this answer

CALCULATE combined with time-intelligence helpers (like DATEADD, SAMEPERIODLASTYEAR, DATESYTD) is the standard DAX pattern for dynamic time calculations that respect the current filter context. CALCULATE modifies the filter context, and the time-intelligence functions generate the appropriate date sets.

Exam trap

DP-600 often tests the misconception that FILTER or LOOKUPVALUE can replace CALCULATE with time-intelligence; the exam expects recognition that CALCULATE plus time-intelligence helpers is the correct context-aware pattern.

How to eliminate wrong answers

Option A is wrong because FILTER with manual date comparisons is error-prone, does not leverage built-in time intelligence, and often breaks under filter context changes. Option C is wrong because ADDCOLUMNS creates calculated columns for a date hierarchy, not a dynamic time-intelligence measure. Option D is wrong because LOOKUPVALUE retrieves a single value by key and cannot perform context-aware time-intelligence calculations like YoY or YTD.

12
Multi-Selecthard

You are designing a semantic model in Microsoft Fabric that will be used by a team of analysts. The model contains several tables and measures. You want to ensure that the model is easy to maintain and that users can quickly find the measures they need. Which two actions should you perform? (Choose two.)

Select 2 answers
A.Create display folders to organize measures by subject area.
B.Add descriptions to measures using the Description property.
C.Rename all tables to include a prefix indicating their source system.
D.Convert all measures to calculated columns to simplify the model.
E.Hide all base columns and only expose measures.
AnswersA, B

Display folders allow you to group measures into logical categories, making it easier for users to navigate and find relevant calculations. This improves discoverability and maintainability, especially in models with many measures. It is a best practice for organizing semantic models.

Why this answer

Creating display folders groups measures logically, and adding descriptions provides context. Together, these actions make measures easier to find and understand, improving both discoverability and maintainability. The other options either do not address the goal or introduce negative side effects.

Exam trap

The trap here is thinking that renaming tables or hiding columns improves organization, but those actions do not directly help users find measures and can hinder self-service.

13
MCQhard

You have a large DirectQuery semantic model and want to implement aggregation tables to improve performance. What is a requirement for the aggregation table to be used by the engine?

A.The aggregation table must be in Import mode.
B.The granularity of the aggregation table must match the visual's request.
C.All columns from the base table must be present in the aggregation table.
D.The aggregation table must contain all historical data rows.
AnswerA

While commonly implemented in Import mode for speed, aggregations can also be implemented in DirectQuery mode. The requirement is not the storage mode itself, but the definition of proper relationships and granularity mapping that allow the engine to substitute the detailed query with the summarized aggregation source.

Why this answer

For an aggregation table to be used by the VertiPaq engine in a DirectQuery model, the aggregation table itself must be stored in Import mode. The engine can then satisfy queries from the in-memory aggregation cache and route only detail-level requests to the DirectQuery source.

Exam trap

DP-600 often tests the misconception that aggregation tables can be DirectQuery or that they must mirror the base table; the key fact is that the aggregation table must be Import mode to be used by the engine.

How to eliminate wrong answers

Option B is wrong because while the aggregation must be at a granularity that can satisfy the query, the requirement for the engine to use it is that it be in Import mode; granularity matching is a design consideration, not the storage-mode requirement. Option C is wrong because aggregation tables intentionally contain a subset of columns (the groupable dimensions and summarizable measures), not all base-table columns. Option D is wrong because aggregation tables typically hold summarized, not all historical detail rows; storing all rows defeats the purpose and is not a requirement.

14
Multi-Selecthard

You are optimizing a slow-performing DAX measure: [Total Profit] = SUMX(Sales, Sales[Quantity] * RELATED(Product[UnitPrice])). The Sales table has 50 million rows. Which THREE actions are recommended to improve the performance of the semantic model and this measure?

Select 3 answers
A.Calculate the 'LineTotal' at the data source or in the Lakehouse and use a simple SUM measure.
B.Change the 'UnitPrice' column in the Product table to a 'Fixed Decimal Number' data type.
C.Use the ALLSELECTED function inside the SUMX to ensure all filters are respected.
D.Ensure the relationship between Sales and Product is based on a surrogate integer key rather than a string key.
E.Wrap the entire SUMX expression in a CALCULATE function to trigger context transition.
AnswersA, B, D

This is the most effective optimization. By calculating 'Quantity * UnitPrice' during the ETL process in the Lakehouse, you create a physical column. The DAX engine can then use a simple SUM function, which is highly optimized and much faster than an iterator like SUMX that has to perform row-by-row math and lookups during report rendering.

Why this answer

Optimizing DAX performance often involves moving logic from query-time to storage-time or simplifying the engine's work. Iterators like SUMX can be slow on large tables. Moving the calculation to a Power Query custom column or a calculated column (though carefully) can help, but the most efficient Fabric approach is performing the calculation in the Lakehouse/Warehouse via SQL or Spark before the data reaches the model.

Exam trap

Candidates often attempt to fix large-scale row-by-row calculations entirely inside DAX using calculated columns or complex iterators, ignoring upstream lakehouse transformations.

15
MCQhard

You are working with a semantic model in Microsoft Fabric that has a measure named 'Total Sales'. You need to create a KPI that compares 'Total Sales' to a target value stored in a table named 'Targets'. The target value is a single scalar value. You want the KPI to display the status (good, bad, neutral) based on thresholds. Which DAX pattern should you use to define the KPI?

A.Create a calculated column in the 'Targets' table that returns the target value, and use it in the KPI visual.
B.Create a measure that uses the LOOKUPVALUE function to retrieve the target from the 'Targets' table, and use it as the target in the KPI visual.
C.Use the KPI visual with a target value that is a constant number, and manually update it when the target changes.
D.Use the KPI visual with a target value from a measure that retrieves the target using SELECTEDVALUE().
AnswerD

The KPI visual in Power BI requires a base measure (e.g., Total Sales) and a target measure. Using SELECTEDVALUE() to retrieve the single target value from the 'Targets' table ensures that the target is correctly applied. This pattern is standard for KPI visuals and allows for dynamic thresholds.

Why this answer

For a KPI visual, the target should be a measure. Using SELECTEDVALUE() to retrieve the target from the 'Targets' table is the correct pattern because it returns a single value and respects filter context. This allows the KPI to dynamically compare the actual measure to the target and display status based on thresholds.

Exam trap

The trap here is assuming that a calculated column or a constant can be used as the target, but the KPI visual specifically requires a measure for the target value.

16
MCQhard

You manage a Microsoft Fabric semantic model that uses Direct Lake mode over a Lakehouse. A report author reports that a measure returns stale data even though new rows were written to the Delta table in the Lakehouse minutes ago. You confirm the Delta table has been updated and the semantic model has no scheduled refresh configured. You need to ensure the model reflects the new data without triggering a full refresh. What should you do?

A.Run a full processing of the semantic model from the Fabric portal.
B.Change the model's storage mode to Import and schedule a refresh.
C.Invoke a refresh of the semantic model, which triggers the default framing of the Delta table.
D.Add a calculated table that queries the Delta table directly.
AnswerC

In Direct Lake mode, a semantic model refresh performs framing of the Delta table, which reads the Delta log and updates the in-memory column segments to include new rows without a full data reload. This is the intended lightweight operation to pick up new Delta commits. It satisfies the requirement because it does not reprocess all partitions from scratch and reflects the newly written rows promptly.

Why this answer

Direct Lake semantic models in Microsoft Fabric read Delta tables through a framing process that maps Delta log metadata to in-memory column segments. When new rows are committed to the Delta table, the semantic model must be refreshed to perform framing and incorporate those rows. This refresh is lightweight compared to a full Import refresh because it does not reload all data; it updates the metadata and column segments to include the new Delta version.

This satisfies the requirement to reflect new data without a full reprocessing of the model.

Exam trap

The trap here is assuming that Direct Lake automatically picks up new Delta commits at query time, when in fact a model refresh is required to perform framing.

17
MCQmedium

You are deploying a semantic model to a Microsoft Fabric workspace. The model contains a measure named [Total Sales] that must be hidden from report authors but still usable in other DAX measures. What should you do?

A.Move the measure to a separate hidden table.
B.Apply object-level security (OLS) to the measure.
C.Set the measure's visibility to hidden in the model.
D.Rename the measure with a leading underscore, such as _Total Sales.
AnswerC

Hiding the measure removes it from the report field list, so report authors cannot accidentally use it directly, but it remains fully accessible to other DAX measures that reference it. This is the standard approach for measures intended only as building blocks within the model, and it does not affect calculation results or model refresh.

Why this answer

Hiding a measure is the correct method to remove it from the report field list while keeping it available for use in other DAX calculations. This ensures report authors cannot use it directly, but model calculations that reference it continue to work. The other options either do not hide the measure or are not applicable to measures.

Exam trap

The trap here is confusing measure visibility with table or column security, which applies to different model objects.

18
MCQmedium

You are designing a semantic model with a complex date requirement. You need to ensure that the model supports multiple relationships between the 'Sales' fact table and the 'Date' dimension (e.g., Order Date, Ship Date). What is the recommended strategy?

A.Create multiple 'Date' dimension tables for each date type.
B.Use the USERELATIONSHIP function in DAX measures.
C.Convert the 'Sales' table into a flat table with redundant date columns.
D.Force all relationships to be bidirectional.
AnswerB

USERELATIONSHIP is the correct and most efficient method to activate an inactive relationship in a specific calculation. It allows you to have a primary date relationship (e.g., Order Date) while still offering the flexibility to pivot data by other dates, such as Ship Date, without creating redundant model objects.

Why this answer

In a star schema, a single Date dimension can be related to the Sales fact table multiple times (Order Date, Ship Date, Due Date), but only one relationship can be active at a time. The USERELATIONSHIP function in DAX allows a measure to activate an inactive relationship for the duration of that measure's evaluation, so you can write separate measures for each date role without duplicating the dimension. This is the standard, recommended pattern in Tabular models.

Exam trap

DP-600 often tests role-playing dimensions — candidates mistakenly believe they need separate Date tables for each date role, when the correct pattern is one dimension with multiple relationships and USERELATIONSHIP in DAX.

How to eliminate wrong answers

Option A is wrong because creating multiple physical Date dimension tables for each date role bloats the model, duplicates data, and complicates maintenance — role-playing dimensions should be handled with one table and multiple relationships. Option C is wrong because flattening the Sales table with redundant date columns denormalizes the model, increases storage, and breaks the star schema design that the VertiPaq engine optimizes for. Option D is wrong because forcing bidirectional cross-filtering on all relationships introduces ambiguity, performance degradation, and unexpected filter propagation — it does not solve the multiple-relationship problem and is considered an anti-pattern.

19
MCQmedium

A large retail company uses Microsoft Fabric and wants to minimize data latency for their executive dashboard. The data resides in Delta tables within a Fabric Lakehouse. You need to choose a storage mode for the semantic model that provides the best performance without requiring data movement or refresh operations. Which mode should you implement?

A.Import mode
B.DirectQuery mode
C.Direct Lake mode
D.Dual mode
AnswerC

Direct Lake mode allows the semantic model to read Parquet files directly from OneLake without a refresh. It bridges the gap between Import and DirectQuery by offering memory-resident performance while maintaining a direct link to the source Delta tables. This ensures that the executive dashboard reflects the most current data available in the Lakehouse.

Why this answer

Direct Lake mode reads Delta tables directly from OneLake using the VertiPaq engine without importing or duplicating data, and without the per-query translation overhead of DirectQuery. Because it leverages the columnar Delta format and the in-memory engine, it delivers Import-like query performance while reflecting near-real-time changes in the lakehouse. This satisfies the requirement for minimal latency with no data movement or scheduled refresh.

Exam trap

DP-600 often tests the distinction between Direct Lake, Import, and DirectQuery — candidates incorrectly assume DirectQuery is the low-latency option because it avoids refresh, when in fact Direct Lake is the only mode that combines no-refresh with Import-class performance.

How to eliminate wrong answers

Option A is wrong because Import mode requires copying data into the semantic model and scheduling refreshes, which introduces latency and data movement — the opposite of the stated requirement. Option B is wrong because DirectQuery translates every visual interaction into a SQL query against the source, causing high latency for complex reports and not providing the best performance for executive dashboards. Option D is wrong because Dual mode is a composite-model feature that lets a table be treated as Import or DirectQuery depending on the query, but it still relies on Import storage and does not eliminate data movement or refresh.

20
MCQhard

You are implementing a large-scale semantic model using Direct Lake. You have a requirement to support a billion-row table while ensuring that historical data is not re-processed every day. You decide to use partitioning. How does partitioning in a Direct Lake model differ from an Import model?

A.Direct Lake partitions must be manually mapped to specific Parquet file paths.
B.Direct Lake does not support partitioning and relies only on V-Order.
C.Partitions in Direct Lake are automatically synced with Delta table partitions.
D.Direct Lake partitions require a SQL gateway to access historical data.
AnswerC

In Direct Lake mode, the semantic model can automatically synchronize its partition structure with the underlying Delta table's physical partitions. This ensures that when the Delta table is updated in the Lakehouse, the semantic model reflects those changes without requiring a traditional data 'import' or 'processing' step.

Why this answer

In Direct Lake mode, partitions are automatically synced with the Delta table partitions in OneLake. When the Delta table is partitioned, the Direct Lake model reflects those partitions without manual mapping, enabling efficient incremental refresh of historical data.

Exam trap

DP-600 often tests the misconception that Direct Lake requires manual partition mapping or a gateway; the key fact is that Direct Lake partitions are automatically synced with Delta table partitions in OneLake.

How to eliminate wrong answers

Option A is wrong because Direct Lake does not require manual mapping to Parquet file paths; it reads Delta tables directly and inherits their partition structure. Option B is wrong because Direct Lake does support partitioning via Delta table partitions; V-Order is a separate optimization, not a replacement for partitioning. Option D is wrong because Direct Lake accesses OneLake directly without a SQL gateway; gateways are for on-premises or certain DirectQuery scenarios, not Direct Lake.

21
MCQmedium

Refer to the exhibit. You are using the Tabular Model Scripting Language (TMSL) to manage a semantic model in a Fabric workspace. What will be the result of executing this command via the XMLA endpoint?

A.The 'Orders_2023' partition in the 'Orders' table will be fully processed and reloaded.
B.The entire 'Orders' table will be refreshed, including all historical partitions.
C.The command will fail because TMSL is not supported for Microsoft Fabric semantic models.
D.Only the metadata for the 'Orders' table will be updated without reloading data.
AnswerA

The TMSL script specifically targets the 'Orders_2023' partition within the 'Orders' table of the 'SalesModel' database. Executing a 'full' refresh type ensures that the engine processes the data for just this segment, allowing for efficient updates of historical data without needing to refresh the entire table.

Why this answer

The exhibit shows a TMSL refresh command targeting a specific partition. By using the 'full' refresh type on a single partition object, the engine will clear the existing data for that specific partition and reload it from the source. This is a common administrative task for managing large datasets where only recent data needs to be updated.

Exam trap

Candidates often confuse partition-level TMSL refresh commands with workspace-level refreshes, mistakenly thinking the entire semantic model is reprocessed when only a specific subset is targeted.

22
MCQeasy

You are developing a semantic model in Microsoft Fabric that will be used by report authors to create paginated reports. The model contains a measure named 'Total Sales' that calculates the sum of sales amounts. You need to ensure that report authors can easily find this measure when building reports. What should you do?

A.Create a perspective that includes only this measure.
B.Set the measure's format string to a currency format.
C.Add a description to the measure that explains its purpose and calculation.
D.Place the measure in a dedicated folder within the model's field list.
AnswerD

Organizing measures into display folders is the primary method to improve discoverability in Power BI. By grouping related measures into folders, you make it easier for report authors to find the measures they need. In Microsoft Fabric, you can create display folders in the model view, and these folders appear in the field list when authoring reports, providing a logical structure.

Why this answer

The most effective way to improve measure discoverability is to organize them into display folders. In Microsoft Fabric, you can create folders in the model view and assign measures to them. This grouping appears in the field list of report authoring tools, making it easier for authors to locate the measures they need.

While descriptions and format strings are useful, they do not aid in finding measures among many. Perspectives can be used for targeted scenarios but are not the primary solution for general discoverability.

Exam trap

The trap here is thinking that adding a description or changing the format string will make measures easier to find, when in fact the field list organization through display folders is what directly impacts discoverability.

23
MCQmedium

You are creating a report for an international company and need to implement field parameters. What is the primary benefit of using this feature for your users?

A.It automatically translates report labels into different languages.
B.It enables users to toggle between different measures or dimensions.
C.It improves the performance of DirectQuery models by caching results.
D.It enforces row-level security across all visual elements.
AnswerB

Field parameters enable users to select which dimensions or measures appear in a visual through a slicer. This flexibility empowers users to customize their view of the data, providing a dynamic experience that replaces multiple static pages and improves the overall usability and interactivity of the report.

Why this answer

Field parameters in Power BI let report authors expose a slicer that switches the fields (measures or dimensions) displayed in a visual at runtime. This gives users self-service flexibility to change what a visual shows — for example, toggling between Sales, Profit, or Quantity — without the author building separate visuals. It is a modeling feature that generates a special table used as a slicer, not a translation or security feature.

Exam trap

DP-600 often tests whether candidates confuse field parameters (switching fields/measures) with calculation groups (switching calculations) or with translation features, so the trap is picking the answer that sounds like localization.

How to eliminate wrong answers

Option A is wrong because label translation is handled by the Q&A linguistic schema or translation files, not field parameters. Option C is wrong because field parameters do not cache DirectQuery results; caching is handled by the Power BI service or Premium capacity features. Option D is wrong because row-level security is enforced by RLS roles defined in the semantic model, not by field parameters.

24
MCQmedium

You are the analytics engineer for a retail company using Microsoft Fabric. The sales team frequently analyzes data by product category and by store location. They want the ability to switch between these two dimensions in a single report page without creating separate visuals. You need to implement a feature that allows users to dynamically change the dimension used in visuals. What should you create?

A.A field parameter
B.A calculation group
C.A measure
D.A hierarchy
AnswerA

Field parameters allow report authors to let users dynamically select which columns or measures are used in a visual. By creating a field parameter containing the product category and store location columns, users can switch between the two dimensions without creating separate visuals or pages, meeting the requirement.

Why this answer

Field parameters enable dynamic selection of columns or measures in a visual, directly addressing the need to switch between product category and store location. This feature is designed for scenario-based dimension switching without creating multiple visuals, making it the correct solution for the sales team's requirement.

Exam trap

The trap here is confusing field parameters with calculation groups; calculation groups modify measures, not the dimensions used for grouping.

25
MCQeasy

You want to improve the discoverability of measures in a semantic model that contains over 50 different calculations. You decide to organize them into logical groups like 'Sales Metrics' and 'Profit Metrics'. Which property should you configure for these measures?

A.Set the 'Display Folder' property for each measure in the Model View.
B.Change the 'Data Category' of the measures to 'Group'.
C.Use the 'Group' feature in the Report View to cluster the measures.
D.Create a new table for each category and move the measures into them.
AnswerA

In the Model View of Power BI Desktop, you can select one or more measures and type a folder name into the 'Display Folder' box. You can even create subfolders by using a backslash (e.g., 'Sales\Current Year'), which helps users find the metrics they need quickly.

Why this answer

The Display Folder property is the correct answer because it allows analytics engineers to organize measures into logical, collapsible groups within the semantic model. When measures are placed in display folders, they appear grouped in the Power BI field list, making it easier for report authors to find related calculations. This is a metadata property that does not affect calculation behavior but significantly improves model usability and discoverability.

Exam trap

DP-600 often tests the confusion between model-level organization features (Display Folder) and report-level features (Group in Report View) — candidates must recognize that measure discoverability in the semantic model is controlled by the Display Folder property, not by report canvas grouping or data categories.

How to eliminate wrong answers

Option B is wrong because 'Data Category' is a property that provides contextual information for a field (such as Geography, City, or Web URL) to help reporting tools interpret the data — it does not create logical groupings and 'Group' is not a valid data category value. Option C is wrong because the 'Group' feature in Report View groups visual elements on a report canvas, not measures in the semantic model — it has no effect on model organization. Option D is wrong because creating separate tables for each measure category is an anti-pattern that fragments the model, complicates relationships, and does not align with tabular modeling best practices — measures should reside in their fact or dedicated measure tables, organized via display folders.

26
MCQmedium

You are managing a Microsoft Fabric semantic model that contains a measure 'Total Sales' defined as SUM(Sales[SalesAmount]). Users frequently run reports that filter by 'Product Category' and 'Year'. You notice that queries are slow when filtering by both attributes simultaneously. You need to improve performance without changing the measure definition. What should you do?

A.Change the measure to use CALCULATE with a filter on Product Category and Year.
B.Create a calculated column that combines Product Category and Year, then use it in slicers.
C.Create an aggregation table that pre-calculates Total Sales by Product Category and Year.
D.Add a composite model with DirectQuery to the underlying data source.
AnswerC

An aggregation table stores pre-aggregated data at a higher grain, such as by Product Category and Year. When users filter by these attributes, the engine can hit the smaller aggregation table instead of scanning the large fact table, dramatically improving query speed. This approach does not require changing the measure definition; the engine automatically uses the aggregation table when possible. It is a recommended optimization in Fabric for frequently queried combinations.

Why this answer

Creating an aggregation table that pre-calculates Total Sales by Product Category and Year is the correct approach. It reduces the data volume scanned when users filter by these attributes, improving query speed without altering the measure. The engine automatically routes queries to the aggregation table when it can satisfy the filter context, making this a seamless optimization.

Exam trap

The trap here is assuming that changing the measure or adding calculated columns will fix performance, when the real solution is to reduce the data scanned via pre-aggregation.

27
MCQmedium

You are designing a semantic model in Microsoft Fabric that will be used by a large number of users. You need to ensure that row-level security (RLS) is enforced for all queries, including those from 'Analyze in Excel'. What should you do?

A.Implement RLS by creating a measure that filters data based on USERPRINCIPALNAME().
B.Configure the 'Analyze in Excel' feature to use a specific service account with limited permissions.
C.Create roles in the semantic model and assign users to them using the Fabric portal.
D.Set the 'Restrict access' property on the semantic model to allow only certain users.
AnswerC

Row-level security in Fabric is implemented by defining roles and assigning users or groups to those roles. When users connect via 'Analyze in Excel', their identity is passed to the model, and RLS is enforced based on their role membership. This is the standard method to secure data at the row level and works across all client tools that support RLS, including Excel.

Why this answer

Row-level security in Microsoft Fabric is enforced by defining roles with DAX filters and assigning users or groups to those roles. This ensures that when users connect via 'Analyze in Excel' or other tools, they only see data permitted by their role. Other methods like measures or service accounts do not provide the same level of security and can lead to data exposure.

Exam trap

The trap here is thinking that dynamic measures or service accounts can enforce RLS, when in fact only properly configured roles with user assignments work across all client tools.

28
MCQmedium

You are developing a Power BI semantic model in Microsoft Fabric. You need to implement row-level security (RLS) to ensure that users in the 'Sales' department can only view data corresponding to their assigned region. Which feature should you use to define these security rules within the semantic model?

A.Object-level security
B.Power BI Deployment Pipelines
C.Manage roles in Power BI Desktop
D.Workspace access settings
AnswerC

Managing roles within Power BI Desktop is the standard method for defining RLS. By creating a role and applying a DAX filter expression to a table, you restrict access to rows. This definition is preserved when the semantic model is published to the Fabric workspace for organizational use.

Why this answer

Row-level security in Fabric semantic models is implemented using security roles defined in the Power BI Desktop model view. By creating roles and using DAX filters, you restrict data access at the row level for specific user groups. This process is essential for compliance and data governance, ensuring that users only interact with information relevant to their business function without needing to create separate reports for every region.

Exam trap

A common mistake is trying to configure row-level security directly inside the Fabric service workspace settings instead of defining the security roles inside Power BI Desktop.

29
MCQeasy

You are a Fabric analytics engineer managing a semantic model that contains a table with columns [OrderDate], [ShipDate], and [DueDate]. You need to create a relationship between this table and a dedicated Date table, but you must ensure that the relationship uses the OrderDate column and that time intelligence functions like TOTALYTD work correctly. What should you do?

A.Create an active relationship between Date[Date] and Sales[OrderDate].
B.Create inactive relationships for all three date columns and use USERELATIONSHIP in every measure.
C.Create an active relationship between Date[Date] and Sales[ShipDate] and mark it as the primary relationship.
D.Create two active relationships: one on OrderDate and one on ShipDate, and let the engine choose based on context.
AnswerA

An active relationship between the Date table and the OrderDate column is required for time intelligence functions to automatically use the correct date column. Since only one relationship between two tables can be active, this ensures that TOTALYTD and similar functions filter Sales based on OrderDate without needing USERELATIONSHIP.

Why this answer

Time intelligence functions in DAX rely on an active relationship between a date table and the fact table to determine the date column to filter. By creating an active relationship on OrderDate, functions like TOTALYTD automatically use OrderDate. The other date columns can have inactive relationships for use with USERELATIONSHIP when needed.

Exam trap

The trap here is assuming that multiple active relationships can exist between two tables or that time intelligence can automatically switch between date columns without an active relationship.

30
MCQmedium

Your organization has a strict data privacy policy that requires certain columns, such as 'Social Security Number', to be completely invisible to unauthorized users, even if they have access to the rest of the semantic model. Which feature should you implement?

A.Object-Level Security (OLS) via Tabular Editor.
B.Row-Level Security (RLS) with a filter on the sensitive column.
C.Sensitivity labels applied at the workspace level.
D.Workspace roles with 'Viewer' permissions only.
AnswerA

OLS is the correct solution for hiding specific columns or tables. While it cannot be configured directly in the Power BI Desktop UI, it can be set up using Tabular Editor by modifying the 'Metadata Protection' property of the column for specific security roles.

Why this answer

Object-Level Security (OLS) in Power BI (implemented via Tabular Editor) restricts access to specific tables or columns, making them completely invisible to unauthorized users. This aligns with the requirement that certain columns like 'Social Security Number' must be hidden entirely, even if users have access to the rest of the model.

Exam trap

DP-600 often tests the confusion between Row-Level Security (RLS) and Object-Level Security (OLS), where candidates might think RLS can hide columns, but it only filters rows.

How to eliminate wrong answers

Option B is wrong because Row-Level Security (RLS) filters rows, not columns; it cannot hide a column entirely. Option C is wrong because sensitivity labels are for classification and protection, not for hiding columns from users. Option D is wrong because workspace roles with 'Viewer' permissions grant read access to the entire semantic model, not column-level hiding.

31
MCQhard

You are designing a semantic model in Microsoft Fabric that will be used by a Power BI report with row-level security (RLS). The model contains a 'Sales' fact table and a 'Customer' dimension table. You need to ensure that when a user from the 'East' region opens the report, they only see sales for customers in the East region. You also want to minimize the impact on query performance. What should you do?

A.Create a role with a DAX filter on the 'Sales' table: RELATED(Customer[Region]) = "East".
B.Create a role with a DAX filter on the 'Sales' table: [Region] = "East".
C.Create a role with a DAX filter on the 'Customer' table: [Region] IN {"East"}.
D.Create a role with a DAX filter on the 'Customer' table: [Region] = "East".
AnswerD

Applying the filter on the dimension table is the most efficient because it leverages the relationship to filter the fact table. The filter propagates from the dimension to the fact, reducing the number of rows processed. This approach aligns with best practices for RLS in Power BI and Fabric, ensuring both security and performance.

Why this answer

Row-level security should be implemented on dimension tables whenever possible, as filters on dimensions propagate to related fact tables efficiently. Filtering the 'Customer' dimension by region ensures that only relevant sales rows are included, and the relationship handles the rest. This approach minimizes performance overhead because the filter is applied to a smaller table and then propagated, rather than evaluating a complex expression on the large fact table.

Exam trap

The trap here is believing that filtering the fact table directly is necessary for RLS, when in fact filtering the dimension table is both correct and more performant.

32
MCQeasy

You are creating a semantic model in Microsoft Fabric. You need to ensure that users can filter a report by product category and see sales data aggregated correctly. The model contains a Sales fact table and a Product dimension table. What should you do?

A.Create a relationship between the Sales table and the Product table using the ProductKey column.
B.Use the LOOKUPVALUE function in a measure to retrieve the product category.
C.Add a calculated column in the Sales table that looks up the product category from the Product table.
D.Create a hierarchy in the Product table that includes Category and Product.
AnswerA

Creating a relationship between the fact table and dimension table on the common key allows filters on the Product dimension to propagate to the Sales fact table. This is the fundamental star schema design for correct aggregation and filtering in Power BI and Fabric semantic models.

Why this answer

To enable filtering of sales data by product attributes, you must create a relationship between the Sales fact table and the Product dimension table. This relationship allows filter context to flow from the dimension to the fact table, ensuring correct aggregation. Other options do not provide dynamic filtering or are not best practices.

Exam trap

The trap here is confusing calculated columns or measures with relationships; only relationships enable dynamic filter propagation between tables.

33
MCQmedium

You are optimizing a large DirectQuery semantic model in Microsoft Fabric. You observe high latency during report interactions due to excessive cross-filtering. Which action should you take to minimize the performance impact while maintaining model integrity?

A.Enable 'Assume referential integrity' for all relationships in the model.
B.Convert all relationships to use bidirectional cross-filtering to ensure data consistency.
C.Restrict relationship cross-filter direction to 'Single' where possible.
D.Increase the 'Query Reduction' setting to disable auto-exist for all columns.
AnswerC

Single-direction filtering limits the path for data propagation, ensuring the engine generates simpler, more efficient SQL queries. By avoiding unnecessary bidirectional relationships, you reduce the workload on the underlying data source, which is crucial for maintaining responsiveness in high-volume DirectQuery semantic models within Fabric.

Why this answer

In a DirectQuery semantic model, bidirectional cross-filtering forces the source database to generate more complex SQL with additional joins and subqueries, which increases latency. Restricting relationships to single-direction cross-filtering where possible reduces the SQL complexity and improves query performance while preserving model integrity. This is the recommended optimization for DirectQuery models with excessive cross-filtering.

Exam trap

DP-600 often tests DirectQuery optimization — candidates incorrectly believe bidirectional filtering improves consistency or performance, when in fact it increases SQL complexity and latency, and the correct action is to restrict to single-direction where possible.

How to eliminate wrong answers

Option A is wrong because 'Assume referential integrity' is an optimization that lets the engine use INNER JOINs instead of OUTER JOINs, but it does not address the cross-filtering overhead and can produce incorrect results if referential integrity is not actually enforced. Option B is wrong because converting all relationships to bidirectional cross-filtering is the opposite of the fix — it increases SQL complexity and latency, and can introduce ambiguous filter paths. Option D is wrong because 'Query Reduction' settings affect how slicers and filters are applied in the client, not the cross-filter direction of relationships, and disabling auto-exist does not reduce the cross-filtering overhead described.

34
MCQeasy

Refer to the exhibit. An analytics engineer needs to connect an external tool to a Microsoft Fabric workspace to manage semantic models. What does the string in the exhibit represent?

A.The OData URL for the Fabric Lakehouse.
B.The connection string for the SQL Analytics Endpoint.
C.The XMLA endpoint for the workspace.
D.The unique identifier for a specific semantic model.
AnswerC

The 'powerbi://' prefix followed by the organization and workspace name is the standard format for a workspace's XMLA endpoint. This allows engineers to connect with tools that support the Analysis Services protocol to perform advanced modeling, scripting, and administrative tasks on semantic models within that workspace.

Why this answer

The XMLA endpoint is the correct answer because it provides a connection point that external tools (such as Excel, Power BI Desktop, Tabular Editor, DAX Studio, and third-party BI tools) use to connect to and manage semantic models in a Microsoft Fabric workspace. The XMLA endpoint URL follows the format 'powerbi://api.powerbi.com/v1.0/myorg/<workspace name>' and supports read/write operations for tabular models. It is the standard protocol for performing management operations like processing, deploying, and modifying semantic models programmatically.

Exam trap

DP-600 often tests whether candidates can distinguish between the XMLA endpoint (used for semantic model management and external tool connectivity) and other Fabric connection strings like the SQL Analytics Endpoint or OData URL — candidates frequently confuse these because they all appear as connection strings in the Fabric UI.

How to eliminate wrong answers

Option A is wrong because the OData URL for a Fabric Lakehouse is used to query Lakehouse data via the OData protocol (typically for OneLake shortcuts and tabular data access), not for managing semantic models. Option B is wrong because the SQL Analytics Endpoint connection string is a TDS/SQL connection string used to query Delta tables in the Lakehouse via T-SQL, not to manage semantic models. Option D is wrong because a unique identifier for a semantic model would be a GUID or object ID, not a connection endpoint string used by external tools.

35
MCQhard

Refer to the exhibit. You are using the XMLA endpoint to manage a large semantic model in a Microsoft Fabric Premium workspace. You execute the TMSL script provided in the exhibit. What is the specific outcome of this operation?

A.It deletes all data in the Orders table except for the 2023 partition.
B.It performs a full data reload for only the specified partition of the Orders table.
C.It recalculates all measures in the model based on the 2023 data only.
D.It updates the metadata for the Orders table without loading any new data.
AnswerB

The JSON script explicitly targets the 'Orders_2023' partition within the 'Orders' table for a 'full' refresh. This operation reloads all data from the source for that specific partition only, allowing for efficient updates to specific time periods without impacting the rest of the model's data or processing time.

Why this answer

The TMSL script shown in the exhibit is a Refresh command targeting the '2023' partition of the 'Orders' table. In TMSL, the Refresh command with Type 'full' on a specific partition performs a full data reload for that partition only, not the entire table. This is a common operation to update a single partition without affecting others.

Exam trap

DP-600 often tests the granularity of TMSL commands; candidates may confuse a partition-level refresh with a table-level or database-level refresh, or assume that a refresh deletes data.

How to eliminate wrong answers

Option A is wrong because the script does not delete data; it refreshes (reloads) the partition. Option C is wrong because the script does not recalculate measures; it only refreshes data for the partition. Option D is wrong because the script does not update metadata; it loads data into the partition.

36
MCQmedium

You are managing a semantic model in Microsoft Fabric that uses Direct Lake mode. The model is based on a Delta table in a lakehouse. Users report that queries are slow when they filter on a column that has a high cardinality. You want to improve performance without changing the model's storage mode. What should you do?

A.Enable query caching on the semantic model.
B.Optimize the Delta table by using Z-Ordering on the high-cardinality column.
C.Add a column-level aggregation to the Delta table.
D.Convert the model to Import mode.
AnswerB

Z-Ordering is a technique in Delta Lake that co-locates related data in the same set of files, improving data skipping and query performance for high-cardinality columns. In Direct Lake mode, since the model reads directly from Delta tables, optimizing the table with Z-Ordering can significantly reduce the amount of data scanned, thus speeding up queries.

Why this answer

In Direct Lake mode, the semantic model queries Delta tables directly. To improve performance for high-cardinality filters, optimizing the Delta table with Z-Ordering on the frequently filtered column allows for better data skipping, reducing the amount of data read and speeding up queries without changing the storage mode.

Exam trap

The trap here is thinking that enabling query caching or converting to Import mode is the solution, but Direct Lake performance is primarily improved by optimizing the underlying Delta table.

37
MCQhard

Refer to the exhibit. You are using the DAX Query View in Power BI Desktop to troubleshoot a performance issue. What is the primary purpose of the 'EVALUATE' statement in this context?

A.It defines a new calculated table that will be permanently added to the semantic model.
B.It triggers the creation of a relationship between the 'Date' and 'Sales' tables.
C.It executes the table expression and returns the result set for inspection.
D.It calculates the total sum of the 'Sales' column for the entire model.
AnswerC

EVALUATE is a required keyword in a DAX query that specifies the table expression to be executed. By using it in DAX Query View, you can verify that your variables and logic (like SUMMARIZECOLUMNS) are producing the correct rows and columns before implementing them in a report.

Why this answer

In DAX Query View, the EVALUATE statement executes a table expression and returns the resulting rows for inspection, making it the core of ad-hoc DAX querying. It is used to test measures, troubleshoot performance, and validate model logic without persisting anything to the model. This is the primary purpose of EVALUATE in that context.

Exam trap

The trap is conflating EVALUATE (a query-time table expression evaluator) with model-modifying operations like calculated tables or relationships — candidates who have not used DAX Query View may assume EVALUATE persists changes.

How to eliminate wrong answers

Option A is wrong because EVALUATE does not create calculated tables — that requires a calculated table definition (e.g., via Tabular Editor or the model designer), not a query. Option B is wrong because relationships are defined in the model schema, not triggered by DAX queries. Option D is wrong because EVALUATE returns a table expression's result set; summing a column requires an explicit aggregation like SUMMARIZECOLUMNS or a measure, not EVALUATE alone.

38
MCQhard

A company has a massive Fact table with 5 years of historical data. They require real-time reporting for the current day's data, but the historical data only needs to be updated once a day. You want to implement a 'Hybrid Table' to meet this requirement. How does a Hybrid Table handle this scenario?

A.The table is stored entirely in Direct Lake mode, which automatically handles both history and real-time.
B.The table uses Import mode for historical partitions and DirectQuery for the latest partition.
C.The table stores only the metadata in the model and fetches all data from the source at runtime.
D.The table uses a Dual storage mode that automatically caches frequently accessed historical rows.
AnswerB

This is the definition of a Hybrid Table. By partitioning the data, the engine can query the compressed, high-performance Import partitions for older data. For the 'today' partition, it sends a DirectQuery to the source system, ensuring that the very latest records are included in the report without waiting for a refresh.

Why this answer

A Hybrid Table in Power BI and Fabric uses Import mode for historical partitions and DirectQuery for the latest partition. This allows real-time reporting on the current day's data via DirectQuery, while historical data is imported for fast query performance and only refreshed once a day. This is the defining characteristic of a Hybrid Table.

Exam trap

DP-600 often tests the confusion between Hybrid Tables and other storage modes like Direct Lake or Dual, and candidates may incorrectly attribute automatic caching or full Direct Lake behavior to Hybrid Tables.

How to eliminate wrong answers

Option A is wrong because Direct Lake mode does not automatically handle both history and real-time; it is a different storage mode that reads Delta tables directly, but Hybrid Tables specifically use Import and DirectQuery. Option C is wrong because storing only metadata and fetching all data at runtime describes a DirectQuery-only table, not a Hybrid Table. Option D is wrong because Dual storage mode is a different concept where a table can act as both Import and DirectQuery depending on the query, but it does not automatically cache historical rows; Hybrid Tables explicitly partition the table with different modes.

39
MCQhard

Refer to the exhibit. You are using the XMLA endpoint to manage a semantic model in a Fabric workspace. A developer executes the Tabular Model Scripting Language (TMSL) command shown. What is the result of this operation?

A.The entire 'SalesModel' database is refreshed, including all tables and partitions.
B.Only the '2023_Q4' partition of the 'Transactions' table is processed and updated.
C.The schema of the 'Transactions' table is updated, but no data is loaded.
D.The command fails because the 'type' property must be set to 'dataOnly' for partitions.
AnswerB

The JSON structure explicitly defines the target as a single partition. By specifying the database, table, and partition, the engine knows exactly which subset of data to reload. This is a common practice for large datasets to save time and resources by only updating the data that has actually changed in the source system.

Why this answer

The TMSL command shown is a `refresh` operation targeting a specific partition, `2023_Q4`, within the `Transactions` table of the `SalesModel` database. The `type` property is set to `full`, which means the partition is completely reprocessed: data is cleared and reloaded, and aggregations are recalculated. Therefore, only that partition is processed and updated, not the entire database or table schema.

Exam trap

DP-600 often tests the misconception that a partition-level refresh with `type: "full"` also updates the table schema or affects the entire database, when in fact it only processes the specified partition's data.

How to eliminate wrong answers

Option A is wrong because the command explicitly scopes the refresh to a single partition (`2023_Q4` of `Transactions`), not the entire database; a database-level refresh would require a different TMSL command without a partition object. Option C is wrong because a `full` refresh processes data and recalculates dependencies; schema updates are not performed by refresh commands—they require separate TMSL `createOrReplace` or `alter` operations. Option D is wrong because the `type` property for a partition refresh can be `full`, `clearValues`, `calculate`, `dataOnly`, `automatic`, or `add`; `dataOnly` is valid but not mandatory, and `full` is perfectly valid.

40
Multi-Selecthard

You are managing a very large semantic model in Microsoft Fabric that exceeds 10 GB in size. You need to ensure that the model can be published and refreshed efficiently. Which TWO configurations are most important for managing large models?

Select 2 answers
A.Enable the 'Auto-scale' feature in the Lakehouse settings.
B.Convert all tables to use the 'Import' mode with 'Basic' compression.
C.Enable 'Large dataset storage format' for the workspace or model.
D.Use the Power BI .PBIX file format for all deployments.
E.Use the XMLA endpoint to perform granular partition refreshes.
AnswersC, E

By default, semantic models have a size limit. Enabling the 'Large dataset storage format' allows the model to grow to the full size of the capacity's memory. This is essential for enterprise-scale models that contain hundreds of millions or billions of rows of data.

Why this answer

Option C is correct because enabling the Large dataset storage format raises the Power BI/Fabric semantic model size limit beyond the default 10 GB (up to 400 GB in Premium/Fabric capacities), which is required for a model that exceeds 10 GB to be published and served. Option E is correct because the XMLA endpoint enables advanced, granular partition management and refresh operations (via TMSL/TOM), so only changed partitions are processed instead of the whole model, which is essential for efficient refresh of very large models. Option A is not relevant because Auto-scale is a Lakehouse/compute setting and does not govern semantic model size or refresh behavior.

Option B is not appropriate because Import mode with Basic compression does not address the 10 GB model limit and Basic compression is less efficient than the default for large models. Option D is not correct because .PBIX is a development/deployment file format and does not enable large-model storage or efficient partition refresh.

Exam trap

DP-600 often tests the misconception that generic capacity features like autoscaling or file format choices solve large-model limits, when the real levers are the large dataset storage format and XMLA-based partition refresh.

41
Multi-Selecthard

You are creating a semantic model that requires complex time-intelligence logic and the ability for users to switch between different currencies. You decide to use calculation groups. Which TWO of the following are true regarding the implementation and behavior of calculation groups in Microsoft Fabric semantic models?

Select 2 answers
A.Calculation groups automatically enable the implicit measure feature for all columns in the model.
B.You must define a precedence property for each calculation group to resolve calculation order.
C.Calculation groups are only supported in models that use the DirectQuery storage mode.
D.The SELECTEDMEASURE function is used within calculation items to reference the measure in context.
E.Each calculation group can only contain a maximum of ten calculation items for performance reasons.
AnswersB, D

Precedence is a mandatory integer property that determines the order in which calculation groups are evaluated. When a single measure is affected by multiple calculation groups, the group with the highest precedence value is evaluated first, which is critical for scenarios involving nested transformations like currency and time intelligence.

Why this answer

Calculation groups allow you to significantly reduce the number of measures by defining reusable logic as calculation items. In Microsoft Fabric, they provide a powerful way to handle multiple transformations, such as YTD or Currency conversion, across various measures. However, they require careful configuration of precedence to resolve conflicts when multiple groups are applied to a single visual.

Exam trap

Candidates forget to set the precedence property when implementing multiple calculation groups, leading to unpredictable evaluation orders and incorrect calculations in visual reports.

42
MCQeasy

You are designing a semantic model in Microsoft Fabric that will be used by report authors to create their own measures. You want to ensure that report authors can easily discover and reuse existing measures without having to search through all tables. What should you do?

A.Create a dedicated measures table and move all measures into it.
B.Use calculation groups to organize measures by category.
C.Set the 'Is hidden' property to false for all measures.
D.Add all measures to the fact table so they appear together.
AnswerA

Placing measures in a dedicated table (often called a measures table or _Measures) groups them together in the field list, making them easier to find and reuse. This practice improves discoverability and organization without affecting the underlying data model. It is a recommended best practice for semantic models that contain many measures, as it separates measures from data columns and reduces clutter.

Why this answer

Creating a dedicated measures table consolidates all measures into a single location in the field list, making them easier to find and reuse. This organizational technique is widely recommended for semantic models with many measures, as it separates measures from data columns and improves the authoring experience for report creators.

Exam trap

The trap here is thinking that changing visibility properties or using calculation groups will organize measures, when the real solution is to physically group them in a dedicated table.

43
MCQhard

You are optimizing a Power BI semantic model that uses Incremental Refresh. You need to ensure that historical data is not reloaded during every refresh cycle. Which configuration is required to achieve this efficiently?

A.Use a fixed date filter in the Power Query source step.
B.Configure 'RangeStart' and 'RangeEnd' parameters as Date/Time types.
C.Enable 'Fast Combine' in the global data load options.
D.Increase the refresh timeout to allow all data to be re-processed.
AnswerB

The incremental refresh engine relies on these specific parameter names and their Date/Time data types to inject filter logic into the source queries. When correctly defined, the engine dynamically updates these parameters to process only the segments of data that fall within the specified range for each refresh.

Why this answer

Incremental refresh requires defining 'RangeStart' and 'RangeEnd' parameters of type Date/Time. These parameters must be used in a filter step within Power Query to enable query folding, which allows the Power BI service to partition the data and only refresh the necessary segments.

Exam trap

Candidates often confuse Incremental Refresh with 'Scheduled Refresh' or assume that simply enabling the feature in the UI is sufficient without configuring the mandatory Power Query parameters.

44
MCQeasy

You are creating a semantic model in Microsoft Fabric that will be used by report authors to build Power BI reports. You need to provide a consistent, reusable calculation for year-over-year sales growth that can be used across multiple reports. What should you create?

A.A calculated column in the Sales table.
B.A Power Query custom column in the data transformation step.
C.A measure using time intelligence functions.
D.A calculated table that stores year-over-year growth.
AnswerC

Measures are calculated at query time and respect filter context, making them ideal for dynamic calculations like year-over-year growth. Time intelligence functions such as SAMEPERIODLASTYEAR or DATEADD can be used to compare sales across periods. Once created, the measure can be reused in any report connected to the semantic model, ensuring consistency.

Why this answer

Measures are the correct choice for dynamic, reusable calculations in a semantic model. They are evaluated at query time and can leverage DAX time intelligence functions to compute year-over-year growth based on the current filter context. This ensures that report authors can use the measure consistently across different reports and time periods.

Exam trap

The trap here is confusing calculated columns or tables with measures, which are the only objects that dynamically respond to filter context.

45
MCQhard

You are optimizing a semantic model in Microsoft Fabric that contains a Direct Lake table. You notice that queries against the model are slower than expected, even though the data is in Delta format. You suspect that the Delta table has many small files. What should you do to improve query performance?

A.Run the OPTIMIZE command on the Delta table to compact small files.
B.Increase the memory allocated to the Fabric capacity.
C.Enable partitioning on the Delta table.
D.Convert the Delta table to Parquet format.
AnswerA

The OPTIMIZE command in Delta Lake compacts small files into larger ones, reducing the number of files that Direct Lake must read. This improves scan performance and reduces query latency. It is a recommended maintenance operation for Delta tables used by Direct Lake.

Why this answer

Small files in Delta tables cause overhead because Direct Lake must open and read each file individually. The OPTIMIZE command compacts these small files into larger ones, reducing the number of file operations and improving scan throughput. This is a key maintenance task for optimizing Direct Lake performance.

Exam trap

The trap here is assuming that increasing capacity or partitioning will solve small file issues, when the direct solution is to compact the files using OPTIMIZE.

46
MCQeasy

You are creating a semantic model in Microsoft Fabric that will be used by report authors. You need to ensure that when they use the 'Analyze in Excel' feature, they can see friendly names for tables and columns instead of technical names from the source. What should you configure?

A.Rename the tables and columns in the semantic model to user-friendly names.
B.Add descriptions to the tables and columns in the model.
C.Set the 'Display folder' property for each column in the model.
D.Create a perspective that includes only the relevant tables and columns.
AnswerA

When users connect to a semantic model via 'Analyze in Excel', they see the table and column names exactly as defined in the model. Renaming them to friendly names improves usability and clarity. This is a common practice to hide technical source names and present business-friendly terminology. The model metadata drives the Excel field list, so renaming is the correct approach.

Why this answer

In Microsoft Fabric semantic models, the names of tables and columns are exposed to client tools like Excel. To show friendly names, you must rename the objects in the model itself. This ensures that when users connect via 'Analyze in Excel', they see the business-friendly terminology.

Other properties like display folders or descriptions do not change the actual names used in Excel.

Exam trap

The trap here is confusing display properties like folders or descriptions with the actual object names that are exposed to Excel.

47
MCQmedium

You are using Git integration in Microsoft Fabric to manage your semantic model's version control. You have made changes to the model's metadata in the workspace. How do you ensure these changes are reflected in your remote Azure DevOps repository?

A.The changes are automatically synced to Azure DevOps in real-time.
B.You must download the .PBIX file and manually upload it to the Git repository.
C.Open the Source Control pane in the workspace and click 'Commit'.
D.Run a 'git push' command from a local terminal connected to the workspace URL.
AnswerC

To synchronize changes, you navigate to the Source Control side-pane in the Fabric workspace. It will list all modified items. You then provide a commit message and click the 'Commit' button, which pushes the metadata changes to the connected Azure DevOps or GitHub branch.

Why this answer

In Microsoft Fabric Git integration, changes made in the workspace are not automatically pushed to the remote repository. You must open the Source Control pane in the workspace and click 'Commit' to stage and commit the changes to the connected Azure DevOps branch. This is the documented workflow for syncing workspace changes to Git.

Exam trap

DP-600 often tests whether candidates know Fabric Git integration requires a manual commit via the Source Control pane — many assume it is automatic like OneDrive sync or that a local git push is needed.

How to eliminate wrong answers

Option A is wrong because Fabric Git integration is not real-time automatic; it requires an explicit commit action. Option B is wrong because Fabric Git integration works with item definitions (e.g., semantic model metadata as TMDL/JSON), not by manually uploading .PBIX files. Option D is wrong because there is no local terminal connected to the workspace URL for git push; the commit is performed through the Fabric UI or APIs, not a local git client.

48
Multi-Selectmedium

You are setting up a deployment pipeline in Microsoft Fabric to move a semantic model from the Development workspace to the Test workspace. Which TWO settings are typically managed using deployment rules during this process? (Choose two.)

Select 2 answers
A.Data source connection strings.
B.The DAX logic of calculated columns.
C.Parameter values defined in Power Query.
D.The Row-Level Security (RLS) role membership.
E.The visual layout of report pages.
AnswersA, C

Deployment rules allow you to redirect a semantic model from a development database to a test or production database automatically. This ensures that the model in the Test workspace points to the validated test data rather than the volatile data used during the initial development phase.

Why this answer

Deployment rules in Microsoft Fabric's deployment pipelines are designed to let you change environment-specific configuration as content moves between stages, and the two settings typically managed this way are data source connection strings (option A) and parameter values defined in Power Query (option C). Option A is correct because a semantic model that points to a Dev SQL server or lakehouse must be redirected to the Test data source when promoted, and a deployment rule for the data source lets you swap that connection string without editing the model. Option C is correct because Power Query parameters (for example, a parameter holding a server name, database name, or file path) commonly differ per stage, and deployment rules allow you to override those parameter values for the Test workspace.

Option B is not managed by deployment rules because calculated column DAX is model logic that should remain identical across stages, not environment-specific configuration. Option D is not a deployment-rule setting because RLS role membership is security configuration managed separately (for example, via workspace roles or the model's security settings), not a stage-specific rule. Option E is not applicable because report page visual layout is report design content that travels with the item and is not an environment-specific setting handled by deployment rules.

Exam trap

DP-600 often tests the misconception that deployment rules can modify model logic (DAX, RLS, visuals); in reality they only swap data source connections and Power Query parameter values.

49
MCQmedium

You are implementing a semantic model in Microsoft Fabric that must show sales amounts in the reporting user's preferred currency. The model contains a fact table 'Sales' with 'SalesAmount' in USD and a dimension table 'Currency' with exchange rates. You need to ensure that when a user slices by currency, the amounts are converted correctly and aggregate properly. Which DAX pattern should you use in a measure?

A.SUMX(Sales, Sales[SalesAmount] * RELATED(Currency[Rate]))
B.SUM(Sales[SalesAmount]) * AVERAGE(Currency[Rate])
C.CALCULATE(SUM(Sales[SalesAmount]), USERELATIONSHIP(Sales[CurrencyKey], Currency[CurrencyKey]))
D.SUMX(Currency, SUM(Sales[SalesAmount]) * Currency[Rate])
AnswerA

This pattern iterates over each row in Sales, multiplies the row's SalesAmount by the related exchange rate from the Currency table, and then sums the results. It ensures that conversion happens at the row level before aggregation, which is required when exchange rates vary by currency and date. RELATED works because there is a relationship from Sales to Currency.

Why this answer

The correct pattern uses row-level iteration with RELATED to fetch the exchange rate for each sale, then multiplies and sums. This ensures that each transaction is converted using its associated currency rate before aggregation. Other approaches either aggregate too early, ignore row context, or rely on relationship functions that do not perform conversion.

Exam trap

The trap here is assuming that multiplying the total sales by an average rate works for multi-currency conversion, but exchange rates must be applied per row.

50
MCQeasy

You need to provide report users with the ability to dynamically change the measure displayed in a bar chart (e.g., switching between 'Total Sales' and 'Total Profit'). Which feature in a Microsoft Fabric semantic model should you use to implement this with minimal effort?

A.Create a separate visual for each measure and use bookmarks to switch between them.
B.Implement Field Parameters for the required measures.
C.Use the SWITCH function in a new DAX measure to select between the two calculations.
D.Configure a drill-through action that navigates to different pages based on the measure.
AnswerB

Field parameters allow you to group multiple measures into a single table. When this table's field is added to a visual, users can use a slicer to select which measure they want to see. This is the most efficient way to provide dynamic measure selection in Power BI reports.

Why this answer

Field Parameters allow report users to dynamically change the measure or dimension used in a visual without creating multiple visuals or complex DAX. By creating a field parameter that includes the measures 'Total Sales' and 'Total Profit', you can add a slicer to the report that lets users switch between them. This is the most efficient and user-friendly approach.

Exam trap

DP-600 often tests the difference between field parameters and other dynamic switching methods like SWITCH measures or bookmarks. The trap is to overcomplicate the solution with DAX when field parameters are the intended minimal-effort feature.

How to eliminate wrong answers

Option A is wrong because creating separate visuals and using bookmarks to switch is cumbersome, requires maintenance, and does not provide a seamless dynamic switch within the same visual. Option C is wrong because while a SWITCH measure can dynamically select a calculation based on a slicer selection, it requires additional DAX and a supporting table, and it is not as straightforward as field parameters; it also may not work well with certain visual behaviors. Option D is wrong because drill-through navigates to a different page and does not change the measure within the same visual; it is not designed for dynamic measure switching.

51
Multi-Selectmedium

You are configuring incremental refresh for a large fact table in a Microsoft Fabric semantic model. Which THREE parameters or settings must be correctly configured to ensure incremental refresh works as expected?

Select 3 answers
A.RangeStart and RangeEnd parameters of type Date/Time.
B.A calculated column that determines the age of each record in days.
C.The 'Archive data starting before' setting in the refresh policy.
D.The 'Enable DirectQuery' checkbox in the table properties.
E.The 'Incrementally refresh data starting before' setting in the refresh policy.
AnswersA, C, E

These two parameters are reserved names used by Power BI to dynamically filter the data source into partitions. They must be set as Date/Time data types. The service uses these to pass the specific date ranges for each partition during the refresh process, ensuring only the necessary data is loaded.

Why this answer

Incremental refresh in a Fabric/Power BI semantic model requires two Date/Time parameters, conventionally named RangeStart and RangeEnd, which define the filter window applied to the table's date column (option A). The refresh policy then uses two policy settings that must be configured: 'Archive data starting before' (option C), which specifies how much historical data is retained and stored, and 'Incrementally refresh data starting before' (option E), which defines the rolling window of data that is refreshed on each incremental run. These three elements together let the engine partition the table by date, refresh only recent partitions, and keep older partitions archived.

Option B is unnecessary because the model uses the RangeStart/RangeEnd parameters and a date column directly, not a calculated age column, to determine partitions. Option D is incorrect because incremental refresh does not require DirectQuery; it works with Import mode (and DirectQuery/Dual tables do not use this partitioning mechanism in the same way).

Exam trap

DP-600 often tests the assumption that incremental refresh requires calculated columns or DirectQuery, when in fact it relies solely on the RangeStart/RangeEnd parameters and the two policy settings.

52
MCQhard

You are optimizing a semantic model in Microsoft Fabric that uses Direct Lake mode. Users report that queries against a large dimension table are slow, especially when filtering by multiple attributes. You examine the Delta table and notice that it is not optimized for query performance. What should you do to improve performance?

A.Increase the Fabric capacity SKU to provide more memory.
B.Convert the dimension table to a calculated table in the semantic model.
C.Enable automatic aggregations on the dimension table.
D.Apply V-Order optimization to the Delta table.
AnswerD

V-Order is a write-time optimization for Delta tables in Microsoft Fabric that sorts and compresses data to accelerate read performance. It is particularly beneficial for Direct Lake models because it reduces the amount of data that needs to be paged into memory. Applying V-Order to the dimension table can significantly improve query speed, especially for filters on multiple attributes.

Why this answer

V-Order is a key optimization for Delta tables used by Direct Lake models in Microsoft Fabric. It sorts and compresses data to improve read efficiency, reducing the amount of data paged into memory and accelerating query performance. Applying V-Order to the dimension table directly addresses the slow query issue without increasing costs or changing the model architecture.

Exam trap

The trap here is to resort to scaling up capacity or changing the storage mode, when the correct fix is to optimize the Delta table with V-Order.

53
MCQmedium

Your team needs to create a semantic model that supports multiple currencies. Instead of creating individual measures for 'Total Sales USD', 'Total Sales EUR', and 'Total Sales GBP', you want to implement a single 'Total Sales' measure that changes its calculation based on a user's selection. Which feature should you use?

A.Field Parameters
B.Calculation Groups
C.Dynamic Format Strings
D.Measure Hierarchies
AnswerB

Calculation Groups allow you to define 'Calculation Items' that act as a layer over your measures. You can create a 'Currency' calculation group where each item (USD, EUR, GBP) applies a specific multiplication factor to the SELECTEDMEASURE(). This allows one 'Total Sales' measure to represent any currency depending on the user's slicer selection on the report.

Why this answer

Calculation Groups are the correct feature for this requirement. They allow for the creation of calculation items that can modify the behavior of existing measures based on a selection in a slicer. While Dynamic Format Strings are often used in conjunction with Calculation Groups to handle currency symbols, the core logic for switching the calculation itself is handled by the Calculation Group items.

Exam trap

Candidates often confuse Calculation Groups with Field Parameters. Field Parameters are used to swap measures or dimensions in a visual, whereas Calculation Groups are used to modify the DAX logic applied to existing measures.

54
Multi-Selecthard

You are managing a Power BI semantic model and need to implement row-level security (RLS) for a global sales team. Which TWO actions must you perform to ensure the security policies are correctly applied and validated?

Select 2 answers
A.Define roles and filter expressions in Power BI Desktop.
B.Assign users to roles using the Security settings in the Fabric workspace.
C.Create separate reports for each regional manager to ensure data isolation.
D.Enable 'View as role' for all workspace members automatically.
E.Apply object-level security (OLS) to all columns containing sensitive data.
AnswersA, B

Roles must be defined within the semantic model using DAX expressions to identify which rows users can access. This logical layer is established during development in Power BI Desktop, allowing developers to create granular security definitions that filter data based on user attributes or organizational hierarchies.

Why this answer

Option A is correct because RLS is authored in Power BI Desktop by creating roles and DAX filter expressions (for example, [Region] = "West") on the semantic model's tables, which defines the actual security logic. Option B is correct because after the model is published, users or groups must be assigned to those roles in the Fabric workspace's semantic model Security settings; without role membership, the filters never take effect for anyone. Option C is not required because RLS is designed to enforce data isolation within a single shared report/model rather than duplicating reports per manager.

Option D is wrong because 'View as role' is a testing feature for validating filters, not something that automatically grants or applies roles to all workspace members. Option E is incorrect here because OLS restricts columns/tables, not rows, and is not a required step for implementing row-level security.

Exam trap

DP-600 often tests the misconception that defining roles in Desktop is sufficient; candidates forget that users must also be assigned to roles in the workspace, or they confuse RLS with OLS.

55
MCQmedium

You are managing a semantic model in Microsoft Fabric that uses Direct Lake mode. The model is connected to a Lakehouse that contains a Delta table with sales data. Users report that queries against the model sometimes return stale data, even though the Delta table has been updated. You need to ensure that the semantic model reflects the latest data as soon as possible. What should you do?

A.Trigger a frame refresh on the semantic model.
B.Set up a scheduled refresh on the semantic model.
C.Increase the memory allocated to the semantic model.
D.Configure the semantic model to use DirectQuery instead of Direct Lake.
AnswerA

In Direct Lake mode, data is loaded into memory via frames that reference the Delta table. When the Delta table is updated, the semantic model may not immediately reflect changes until a frame refresh occurs. Triggering a frame refresh (e.g., via XMLA endpoint or REST API) forces the model to reload the latest data, ensuring users see up-to-date information.

Why this answer

Direct Lake models use frames to load data from Delta tables. When the underlying Delta table changes, the model may not automatically reflect those changes until a frame refresh is performed. Triggering a frame refresh ensures the latest data is loaded into memory, resolving the staleness issue.

Exam trap

The trap here is assuming that Direct Lake automatically stays in sync with Delta table updates or that scheduled refresh applies; frame refresh is the correct mechanism.

56
MCQmedium

Refer to the exhibit. You are configuring dynamic format strings for a currency column. Based on the provided JSON policy configuration, which outcome will occur when a user views the report?

A.The report will fail to render because the JSON policy is invalid.
B.The column will display only the US format string for all rows.
C.The display format will change based on the filtered country context.
D.The report will ignore the policy and use the default model format.
AnswerC

Dynamic format strings allow the formatting property of a measure to be determined by a DAX expression. As the filter context changes, the engine re-evaluates the expression, causing the currency symbols to toggle between USD and EUR based on the logic defined within the model metadata.

Why this answer

The dynamic format string evaluates the DAX expression for every data point displayed in the visual. By checking the country context, the engine applies either the USD or EUR format string dynamically. This feature is vital for global reports, as it allows a single measure to adapt its visual representation based on user context, improving clarity and usability without creating separate measures for each currency type.

Exam trap

Candidates often assume the format string is applied at the model level globally. They overlook that dynamic strings evaluate per data point based on the current filter context of the visual.

57
MCQmedium

You are managing a semantic model in Microsoft Fabric that has a large Sales fact table. Users report that queries filtering on a specific product category are slow. You notice that the Product table has a relationship to Sales, and the ProductCategory column is frequently used in filters. What should you do to improve query performance?

A.Set the ProductCategory column to use the 'Sort by column' property.
B.Create a user hierarchy in the Product table that includes ProductCategory.
C.Create an aggregation table that pre-calculates sales by ProductCategory.
D.Add a relationship between ProductCategory and Sales.
AnswerC

Aggregation tables pre-calculate and store summarized data, which can significantly improve query performance for common aggregations and filters. By creating an aggregation table that summarizes sales by ProductCategory, queries that filter on ProductCategory can hit the smaller aggregation table instead of the large fact table, reducing the amount of data scanned.

Why this answer

Creating an aggregation table that pre-calculates sales by ProductCategory can dramatically improve query performance. Aggregations reduce the data volume that needs to be scanned for common queries, allowing the engine to satisfy queries from the smaller aggregated table. This is especially effective when users frequently filter on ProductCategory and aggregate sales measures.

Exam trap

The trap here is assuming that adding a hierarchy or changing sort properties will improve performance, when in fact aggregations are designed specifically to optimize query performance for frequently used aggregations and filters.

58
MCQeasy

You are the analytics engineer for a retail company using Microsoft Fabric. A semantic model contains a Sales fact table and a Date dimension table. Report authors frequently filter by year and quarter but complain that the Date table is not recognized as a date table, causing time intelligence functions to return incorrect results. What should you do to resolve this?

A.Mark the Date table as a date table in the semantic model.
B.Change the data type of the Date column in the Date table to Text.
C.Create a relationship between the Sales table and the Date table on the OrderDate column.
D.Add a calculated column to the Sales table that extracts the year from the OrderDate column.
AnswerA

Marking the Date table as a date table ensures that time intelligence functions like TOTALYTD and SAMEPERIODLASTYEAR work correctly. It also allows the use of the Date table's hierarchy for filtering. This is the standard fix when time intelligence returns incorrect results due to the Date table not being recognized.

Why this answer

Marking the Date table as a date table is required for time intelligence functions to work correctly. It ensures that the table contains a continuous set of dates and that the relationship to the fact table is properly utilized. Without this designation, functions like TOTALYTD may return inaccurate results, and the Date table won't be used for filtering in the expected way.

Exam trap

The trap here is assuming that creating a relationship alone is sufficient for time intelligence, when in fact marking the table as a date table is a separate required step.

59
MCQmedium

You are designing a semantic model in Microsoft Fabric that will use Direct Lake mode. The underlying data is stored in a Lakehouse as Delta tables. You need to ensure the highest performance for report queries while maintaining the model in Direct Lake mode. Which data modeling strategy should you prioritize to achieve this goal?

A.Flatten all dimension and fact data into a single wide table to eliminate all joins.
B.Implement a snowflake schema to ensure that data normalization reduces the overall storage size.
C.Organize the data into a star schema with clearly defined fact and dimension tables.
D.Use many-to-many relationships between all tables to provide maximum flexibility for users.
AnswerC

A star schema is the recommended architecture for semantic models in Fabric because it optimizes how the engine handles relationships. In Direct Lake mode, this structure allows for faster column scanning and simpler join paths, which keeps the model in its performant state and prevents unnecessary fallback to DirectQuery for complex operations.

Why this answer

Direct Lake mode provides the performance of Import mode by loading Parquet files directly from OneLake into memory without duplicating data or requiring a refresh. To maintain this performance, models should adhere to a star schema. This reduces join complexity that the engine must resolve during query execution, ensuring the VertiPaq engine can efficiently scan and compress the data columns retrieved from the Delta tables.

Exam trap

Candidates often suggest 'denormalizing' all data into a single flat table. While this seems fast, a star schema is the required architecture for optimal Direct Lake performance.

60
Multi-Selectmedium

You are implementing security for a Microsoft Fabric semantic model. You need to ensure that regional managers can only see data for their own regions. Which TWO steps must you perform to successfully implement and validate Dynamic Row-Level Security (RLS)?

Select 2 answers
A.Define a DAX filter expression using the USERPRINCIPALNAME function in a model role.
B.Enable the 'Allow users to modify data' setting in the workspace permissions.
C.Assign users or security groups to the defined role within the Power BI service.
D.Create a separate semantic model for each regional manager.
E.Configure Object-Level Security (OLS) to hide the 'Region' column from all users.
AnswersA, C

This is the technical core of dynamic RLS. By using USERPRINCIPALNAME() in a role's DAX expression, the model can compare the logged-in user's email address against a column in the security table. This ensures that the data is filtered automatically based on the identity of the person accessing the report at that specific moment.

Why this answer

Option A is correct because Dynamic RLS in a Fabric/Power BI semantic model is implemented by creating a model role and defining a DAX filter expression on the relevant table (for example, [Region] = LOOKUPVALUE(...) or a filter using USERPRINCIPALNAME() to map the signed-in user to their region), which is the mechanism that dynamically restricts rows per user. Option C is correct because the role only takes effect once users or security groups are assigned to it in the Power BI service (via the semantic model's Security settings), which is also how you validate that each regional manager sees only their own region's data. Option B is incorrect because 'Allow users to modify data' relates to write permissions in the workspace and has nothing to do with filtering rows for RLS.

Option D is incorrect because creating a separate semantic model per manager is a duplication approach, not Dynamic RLS, and defeats the purpose of a single secured model. Option E is incorrect because OLS hides columns or tables, not rows, so it cannot restrict regional managers to their own region's rows.

Exam trap

DP-600 often tests the steps for RLS, and candidates may forget that assigning users to roles in the service is required. The trap is to focus only on the DAX and overlook the assignment step.

61
MCQhard

Refer to the exhibit. You are writing a DAX measure that compares a Sales ID from a fact table to a numeric ID in a dimension table. Which action will resolve this error?

A.Use the FORMAT function on the integer column.
B.Change the data type of the ID columns to be consistent.
C.Wrap the comparison in an IFERROR function.
D.Disable strict type checking in the model settings.
AnswerB

Ensuring data type consistency is the fundamental requirement for valid DAX comparisons. By aligning the columns to the same type, you eliminate the mismatch, allowing the engine to perform accurate join operations and logical comparisons, which is necessary for correct measure calculation and overall report integrity.

Why this answer

Data type mismatch is a common source of DAX errors. The engine enforces strict type checking during comparisons. You must ensure that both columns share a compatible data type—usually integer for keys—to allow for successful join evaluation.

Converting the column types in Power Query or using DAX conversion functions ensures that the engine can perform efficient joins and comparisons without runtime failure.

Exam trap

Candidates frequently attempt to use DAX functions like FORMAT or SWITCH to force a comparison between strings and integers, which fails due to strict type checking in the VertiPaq engine.

62
Multi-Selecthard

You are designing a semantic model in Microsoft Fabric that will be used by a large number of report authors. The model contains several tables and many measures. You want to improve the user experience by making it easier for authors to find relevant fields and measures. Which two actions should you perform? (Choose two.)

Select 2 answers
A.Create display folders to organize measures and columns into logical groups.
B.Add a description to every measure and column to explain its purpose.
C.Rename all tables and columns to use a consistent naming convention with prefixes.
D.Create perspectives that include only the tables and measures relevant to specific report author roles.
E.Set the 'Is Hidden' property to true for all columns that are not used in reports.
AnswersA, D

Display folders allow you to group related measures and columns together in the field list. This makes it much easier for report authors to navigate a large model and find the fields they need. In Microsoft Fabric, you can create folders in the model view and assign objects to them. This is a best practice for improving discoverability and is a key feature for managing complex semantic models.

Why this answer

To improve discoverability in a large semantic model, you should organize fields into display folders and create perspectives for specific audiences. Display folders group related items, making the field list more navigable. Perspectives filter the model to show only relevant objects for a given role, reducing clutter and focusing authors on what they need.

Both features are supported in Microsoft Fabric and are essential for managing complex models. While descriptions and naming conventions have their place, they do not provide the same level of direct organizational benefit.

Exam trap

The trap here is focusing on documentation (descriptions) or renaming (naming conventions) as the primary means to improve discoverability, when the real solutions are organizational features like display folders and perspectives.

Ready to test yourself?

Try a timed practice session using only Implement Manage Semantic Models questions.