How Should Contoso Access Amazon S3 Email Data in Fabric?
Case Study This is a case study. Case studies are not timed separately. You can use as much exam time as you would like to complete each case. However, there may be additional case studies and sections on this exam. You must manage your time to ensure that you are able to complete all questions included on this exam in the time provided. To answer the questions included in a case study, you will need to reference information that is provided in the case study. Case studies might contain exhibits and other resources that provide more information about the scenario that is described in the case study. Each question is independent of the other questions in this case study. At the end of this case study, a review screen will appear. This screen allows you to review your answers and to make changes before you move to the next section of the exam. After you begin a new section, you cannot return to this section. To start the case study To display the first question in this case study, click the Next button. Use the buttons in the left pane to explore the content of the case study before you answer the questions. Clicking these buttons displays information such as business requirements, existing environment, and problem statements. If the case study has an All Information tab, note that the information displayed is identical to the information displayed on the subsequent tabs. When you are ready to answer a question, click the Question button to return to the question. Overview. Company Overview Contoso, Ltd. is an online retail company that wants to modernize its analytics platform by moving to Fabric. The company plans to begin using Fabric for marketing analytics. Overview. IT Structure The company’s IT department has a team of data analysts and a team of data engineers that use analytics systems. The data engineers perform the ingestion, transformation, and loading of data. They prefer to use Python or SQL to transform the data. The data analysts query data and create semantic models and reports. They are qualified to write queries in Power Query and T-SQL. Existing Environment. Fabric Contoso has an F64 capacity named Cap1. All Fabric users are allowed to create items. Contoso has two workspaces named WorkspaceA and WorkspaceB that currently use Pro license mode. Existing Environment. Source Systems Contoso has a point of sale (POS) system named POS1 that uses an instance of SQL Server on Azure Virtual Machines in the same Microsoft Entra tenant as Fabric. The host virtual machine is on a private virtual network that has public access blocked. POS1 contains all the sales transactions that were processed on the company’s website. The company has a software as a service (SaaS) online marketing app named MAR1. MAR1 has seven entities. The entities contain data that relates to email open rates and interaction rates, as well as website interactions. The data can be exported from MAR1 by calling REST APIs. Each entity has a different endpoint. Contoso has been using MAR1 for one year. Data from prior years is stored in Parquet files in an Amazon Simple Storage Service (Amazon S3) bucket. There are 12 files that range in size from 300 MB to 900 MB and relate to email interactions. Existing Environment. Product Data POS1 contains a product list and related data. The data comes from the following three tables: Products ProductCategories ProductSubcategories In the data, products are related to product subcategories, and subcategories are related to product categories. Existing Environment. Azure Contoso has a Microsoft Entra tenant that has the following mail-enabled security groups: DataAnalysts: Contains the data analysts DataEngineers: Contains the data engineers Contoso has an Azure subscription. The company has an existing Azure DevOps organization and creates a new project for repositories that relate to Fabric. Existing Environment. User Problems The VP of marketing at Contoso requires analysis on the effectiveness of different types of email content. It typically takes a week to manually compile and analyze the data. Contoso wants to reduce the time to less than one day by using Fabric. The data engineering team has successfully exported data from MAR1. The team experiences transient connectivity errors, which causes the data exports to fail. Requirements. Planned Changes Contoso plans to create the following two lakehouses: Lakehouse1: Will store both raw and cleansed data from the sources Lakehouse2: Will serve data in a dimensional model to users for analytical queries Additional items will be added to facilitate data ingestion and transformation. Contoso plans to use Azure Repos for source control in Fabric. Requirements. Technical Requirements The new lakehouses must follow a medallion architecture by using the following three layers: bronze, silver, and gold. There will be extensive data cleansing required to populate the MAR1 data in the silver layer, including deduplication, the handling of missing values, and the standardizing of capitalization. Each layer must be fully populated before moving on to the next layer. If any step in populating the lakehouses fails, an email must be sent to the data engineers. Data imports must run simultaneously, when possible. The use of email data from the Amazon S3 bucket must meet the following requirements: Minimize egress costs associated with cross-cloud data access. Prevent saving a copy of the raw data in the lakehouses. Items that relate to data ingestion must meet the following requirements: The items must be source controlled alongside other workspace items. Ingested data must land in the bronze layer of Lakehouse1 in the Delta format. No changes other than changes to the file formats must be implemented before the data lands in the bronze layer. Development effort must be minimized and a built-in connection must be used to import the source data. In the event of a connectivity error, the ingestion processes must attempt the connection again. Lakehouses, data pipelines, and notebooks must be stored in WorkspaceA. Semantic models, reports, and dataflows must be stored in WorkspaceB. Once a week, old files that are no longer referenced by a Delta table log must be removed. Requirements. Data Transformation In the POS1 product data, ProductID values are unique. The product dimension in the gold layer must include only active products from product list. Active products are identified by an IsActive value of 1. Some product categories and subcategories are NOT assigned to any product. They are NOT analytically relevant and must be omitted from the product dimension in the gold layer. Requirements. Data Security Security in Fabric must meet the following requirements: The data engineers must have read and write access to all the lakehouses, including the underlying files. The data analysts must only have read access to the Delta tables in the gold layer. The data analysts must NOT have access to the data in the bronze and silver layers. The data engineers must be able to commit changes to source control in WorkspaceA. You need to ensure that usage of the data in the Amazon S3 bucket meets the technical requirements. What should you do?
Community Insight
Two constraints must be satisfied simultaneously — no raw-data copies in the lakehouse and reduced cross-cloud egress — and the trap is believing that shortcut caching violates the 'no copy' requirement.
This DP-700 case study tests cross-cloud access to a year of historical email interaction data stored in 12 Parquet files in Amazon S3, where Fabric must both minimize egress costs and avoid storing raw copies in the lakehouses. An OneLake shortcut with caching enabled satisfies both: the raw files remain referenced in S3 rather than landing in bronze, silver, or gold, while cached reads stop repeated cross-cloud transfers.
Option B (shortcut with caching disabled) is the common wrong pick. Learners reason that any caching equals a stored copy, but the cached blocks live in Fabric-managed cache outside the bronze, silver, and gold folders, so the lakehouse stays copy-free while disabling caching forces every read back to S3 and increases egress.
Community Discussion (31 comments)
Comments & Corrections
No comments yet — spotted an error or have a note? Share it below.
Expert Analysis
Why the Answer Is Correct
A OneLake shortcut only references the Amazon S3 Parquet files instead of ingesting them, so no raw copy is persisted in Lakehouse1's bronze, silver, or gold layers — directly meeting the "prevent saving a copy of the raw data" requirement. Enabling caching lets Fabric serve repeated reads from its managed cache rather than re-pulling the 300–900 MB files from S3 on every query, which is precisely how cross-cloud egress is minimized. Because caching is a shortcut/workspace-managed store and not part of the Delta table folders, the raw-data requirement is not violated. This combination is the only option in the list that addresses both stated requirements at once.Why the Other Options Are Wrong
Option B creates the shortcut but disables caching, so every read of the S3 Parquet files crosses the cloud boundary and generates fresh egress charges — this fails the cost-minimization requirement. Options A and C rely on a workspace identity, which is an authentication mechanism (used for example to reach Key Vault-backed secrets or to run Fabric items under a managed identity); it neither prevents raw copies nor reduces data transfer volume from S3. Option A additionally adds high-concurrency notebooks, which governs compute parallelism, not data residency or egress. None of A, B, or C addresses the egress requirement the way shortcut caching does.Community Comment Notes
Most commenters converged on D, with GHill1982 noting that "Enabling caching for the workspace will help minimize egress costs" while a shortcut "ensures that the raw data is not duplicated in the lakehouse." DarkDerf and kim32 argued for B on the grounds that "Caching would store data in Fabric, violating the 'no raw data copies' rule," but as DataAITraining countered, "the requirement was to prevent storing data in the Lakehouse, not Fabric." 670dd08 spelled out the trade-off precisely: without cache, "Data is always read directly from S3" with "high egress costs," whereas the cached shortcut stores data "in a temporary Fabric-managed cache (not in your lakehouse bronze/silver/gold folders)."Official Reference
Exam Strategy
On Fabric loading-pattern questions, always match each option against every stated constraint separately before choosing; here the winning option had to satisfy both 'no raw copy' and 'lower egress.' Remember that OneLake shortcut caching lives in Fabric-managed storage, not in your Delta table folders, so it does not count as a lakehouse copy.
Frequently Asked Questions
Why doesn't shortcut caching break the 'no raw data copies' requirement?
Caching stores data in Fabric-managed cache, not in the bronze, silver, or gold folders of Lakehouse1, so the lakehouse itself holds no raw copy while egress drops sharply.
Why are the workspace identity options A and C incorrect?
A workspace identity handles authentication to services such as Key Vault; it neither avoids copying S3 files nor reduces cross-cloud data transfer, so it satisfies neither requirement.