Task 1.1: Plan a data analytics environment

• AF32

  Incorrect

  F2 to F32 allow 300 million rows per table; with Direct Lake on OneLake the refresh then fails and the model can't be queried.

• BF64

  Correct

  F64 allows 1,500 million rows and 5,000 Parquet files per table and 25 GB of memory per model, so the table frames and its columns fit.

• CF128

  Incorrect

  F128 also meets every limit, but F64 already does, so it isn't the smallest SKU that works.

• DF32, with the tables switched to Direct Lake on SQL so that queries over the large table can fall back

  Incorrect

  Direct Lake on SQL falls back to DirectQuery when a table exceeds the guardrails, which the requirement rules out.

Direct Lake guardrails depend on the SKU. Up to F32, a table may hold 300 million rows and 1,000 Parquet files or row groups, with 10 GB of max memory on F32; F64 raises these to 1,500 million rows, 5,000 files and row groups, and 25 GB. When a table exceeds a guardrail, Direct Lake on OneLake behaves like Import mode: refresh fails and the model can't be queried until the tables fit. Direct Lake on SQL instead falls back to DirectQuery if fallback is enabled.

• AKeep the Pro workspace and refresh through a virtual network data gateway in that network

  Incorrect

  VNet data gateways need a Fabric capacity or a Premium capacity (A4 and higher or any P SKU), which a Pro workspace in shared capacity doesn't provide.

• BMove the workspace to the F16 capacity and refresh through a virtual network data gateway

  Correct

  A VNet data gateway is managed by Microsoft, reaches the database through the private endpoint, and is available on Fabric and Premium capacities.

• CInstall an on-premises data gateway in standard mode on a virtual machine in that network

  Incorrect

  It would reach the private endpoint, but it's a Windows application on a machine that the infrastructure team would have to run and patch.

• DMove the workspace to the F16 capacity and create a managed private endpoint to the database

  Incorrect

  Managed private endpoints serve Fabric Data Engineering items such as notebooks, lakehouses and Spark job definitions, not semantic model refresh.

A virtual network data gateway is injected into a delegated subnet of your Azure virtual network, so it can reach data sources through private endpoints without any traffic on public endpoints, and its hardware is fully managed. It supports Power BI semantic models, paginated reports, Dataflow Gen2, pipelines, Copy job and mirroring, and it requires a Fabric capacity (any F SKU) or a Premium capacity. An on-premises data gateway also works but has to be installed and maintained on a machine.

• AdataColors

  Incorrect

  dataColors is optional; it's only needed when you want to change the series colors.

• Bname

  Correct

  At its most basic level, a theme file needs only the name line; every other property is optional.

• CvisualStyles

  Incorrect

  visualStyles is optional; a theme file is valid without it.

• DtextClasses

  Incorrect

  textClasses is optional; a theme file is valid without it.

A report theme JSON file has only one required line: name. Everything else, such as dataColors, textClasses and visualStyles, is optional, so you add only the properties you want to format and keep Power BI's defaults in the base theme for the rest. Power BI validates the file against a published JSON schema when you import it.

• AA Fabric capacity with an F SKU

  Correct

  Pause and resume is listed as a feature of F SKUs, which Azure bills per second.

• BA Power BI Premium capacity with a P SKU

  Incorrect

  P SKUs don't support pause and resume; they come with a monthly commitment.

• CA Fabric trial capacity

  Incorrect

  A trial capacity doesn't support the features in the SKU parity list unless noted, and pause isn't one of the exceptions.

• DA Premium Per User workspace

  Incorrect

  PPU is a per-user license, not a capacity that you can pause.

Fabric features differ by SKU type. Pause and resume, on-demand resizing, ARM APIs and Terraform, managed private endpoints and trusted workspace access are available on F SKUs and not on P SKUs, while Power BI autoscale is a P SKU feature. F capacities are billed per second through Azure, and pausing one stops the billing until you resume it.

• AUsers can access data stored in OneLake with apps external to Fabric

  Incorrect

  That setting controls external apps such as OneLake File Explorer and ADLS APIs, not exports from semantic models.

• BUsers can create Fabric items

  Incorrect

  It governs creating Fabric items; it doesn't let semantic models write their tables to OneLake.

• CSemantic models can export data to OneLake

  Correct

  This setting lets semantic models configured for OneLake integration send their import tables to OneLake.

• DAllow XMLA endpoints and Analyze in Excel with on-premises semantic models

  Incorrect

  It enables XMLA and Analyze in Excel connections, not OneLake integration.

OneLake integration for semantic models writes the data of import tables to Delta tables in OneLake, where lakehouses, warehouses and notebooks can use them. A Fabric administrator must enable the tenant setting Semantic models can export data to OneLake; model owners then turn on OneLake integration for their models.

• AHer existing personal-mode gateway

  Incorrect

  A gateway in personal mode can be used only with Power BI and supports a single user.

• BA virtual network data gateway in an Azure virtual network

  Incorrect

  It runs in an Azure virtual network, which has no path to a database on the company network without a network link.

• CA managed private endpoint from the workspace

  Incorrect

  Managed private endpoints serve Data Engineering items such as notebooks, not Dataflow Gen2.

• DAn on-premises data gateway in standard mode

  Correct

  Standard mode supports Fabric workloads such as Dataflow Gen2 and centrally managed, shared connections.

The on-premises data gateway runs on a machine in your network and works with Power BI, Microsoft Fabric, Power Apps and other services. Personal mode allows one user to connect to data sources, can't be shared and can be used only with Power BI. Standard mode supports centrally managed connections for multiple users and is the mode Fabric items such as Dataflow Gen2 and pipelines use.