All 33 lessons
Start here: cloud and data basics
- 1. Start here: the exam, the plan and what you already know
- 2. Cloud in plain words: SaaS, PaaS, IaaS and where Fabric fits
- 3. Data basics: files, tables, formats and the lakehouse idea
- 4. A tour of Microsoft Fabric: OneLake, workspaces and items
- 5. The three languages: SQL, PySpark and KQL side by side
- 6. Getting data in: the map of Fabric tools
Store, ingest and transform batch data (exam area 2, 30-35%)
- 7. Choosing a data store and a transformation tool
- 8. Shortcuts and mirroring: use data without copying it
- 9. Ingest data with pipelines and Copy job
- 10. Transform with PySpark and SQL: join, denormalize, aggregate
- 11. Messy data: duplicates, missing values and late arrivals
- 12. Full and incremental loads, and loading a star schema
Orchestrate and stream (exam areas 1 and 2)
Set up and secure a Fabric solution (exam area 1, 30-35%)
- 18. Workspace settings: Spark, domains, OneLake and data workflows
- 19. Version control with Git and database projects
- 20. Deployment pipelines: development, test and production
- 21. Who can see what: workspace roles and item permissions
- 22. Fine-grained security: rows, columns, objects, files and masking
- 23. Governance: sensitivity labels and endorsement
Monitor, fix and optimize (exam area 3, 30-35%)
- 24. Monitoring hub, alerts and Activator
- 25. Fixing errors in pipelines, dataflows and notebooks
- 26. Fixing errors in Eventhouse, Eventstream and T-SQL
- 27. Optimize lakehouse tables: V-Order, OPTIMIZE, VACUUM and partitions
- 28. Optimize Spark, pipelines and warehouses
- 29. Optimize eventstreams, eventhouses and queries
Exam practice and final review
18 min
A tour of Microsoft Fabric: OneLake, workspaces and items
Name every Fabric experience and item a data engineer uses and how they share one copy of data.
By the end of this lesson you can
- Explain what OneLake is and why each tenant has exactly one
- Place tenant, capacity, domain, workspace and item in the right order
- Tell the items that store data from the items that move and transform it
- Say how Power BI reads lake data with Direct Lake
From earlier lessons · 1 of 2
In Fabric, what do you pay for to get compute power?
Pick one answer.
From earlier lessons · 2 of 2
What does Delta Lake add on top of Parquet files?
Pick one answer.
It's your first day as a data engineer at Northwind Rides. You open Fabric and see a dozen item types: Lakehouse, Warehouse, Eventhouse, Eventstream, Notebook and more. Your manager asks where the scooter trip events should land. You can't answer until you know what each box is for.
OneLake: one lake for the whole organization
Exam depth: know it cold.
OneLake is the single data lake that every Fabric tenant gets automatically. A tenant is your organization's one instance of Fabric.
There is exactly one OneLake per tenant. You can't create a second one or delete it, and there's nothing to set up. Teams still keep their own areas inside it, so one lake doesn't mean one big mess.
OneLake stores tables in the open Delta Parquet format. Every Fabric engine reads and writes that same format. So a table written by Spark can be read by SQL or Power BI with no copy and no export. That's the core idea: one copy of data, many engines.
Tenant, capacity, workspace, item
Exam depth: know it cold.
Fabric organizes everything in four levels.
- Tenant: your whole organization. It owns the one OneLake.
- Capacity: a pool of compute inside the tenant, in one region. A tenant can have several.
- Workspace: a container where a team builds and shares work. Each workspace is assigned to one capacity. It also controls who can access what's inside.
- Item: one object you create in a workspace, such as a Lakehouse or a Notebook.
A domain sits beside this chain: it groups workspaces by business area.
A workspace holds up to 1,000 items. Each workspace shows up as a folder in OneLake, and each item as a folder inside it.
Items that store data
Exam depth: know it cold.
Four items hold data. All of them keep it in OneLake.
- Lakehouse: files and Delta tables in one place. You write to it with Spark, an engine that splits big jobs across many machines. Fabric adds a SQL analytics endpoint to every Lakehouse. That endpoint lets you read its tables with T-SQL, Microsoft's dialect of SQL.
- Warehouse: structured tables with full T-SQL. You can insert, update and delete, and run transactions across several tables.
- Eventhouse: a store for streaming, time-stamped events, such as logs and sensor readings. It holds one or more KQL databases, which you query with KQL (Kusto Query Language).
- Mirrored database: a copy of an outside database that Fabric keeps in sync for you. It also gets a SQL analytics endpoint, and it's read-only.
| Need | Pick |
|---|---|
| Spark, files, any shape of data. | Lakehouse |
| SQL-first team, multi-table transactions. | Warehouse |
| Streaming events, queries in seconds. | Eventhouse |
| Live copy of an operational database. | Mirrored database |
Decision rule: pick by the team's language and the data's shape: Spark means Lakehouse, T-SQL writes mean Warehouse, and events mean Eventhouse.
You are a data engineer at Northwind Rides. The finance analysts know only T-SQL. They must load payment data into fact and dimension tables and update several tables in one transaction. You need to choose where they store this data. What should you use?
Pick one answer.
Items that move and transform data
Exam depth: know it cold.
Six items move or reshape data. None of them is where data lives.
- Pipeline (older pages say Data pipeline): runs a series of steps on a schedule or when an event happens, with loops and conditions. It's the conductor.
- Copy job: copies data from a source to a destination, in full or only what changed. No pipeline needed.
- Dataflow Gen2: cleans and reshapes data in a low-code, point-and-click editor called Power Query.
- Notebook: an interactive page where you write and run code, usually PySpark (Python on Spark).
- Spark job definition: runs a finished Spark program as a batch or streaming job. It needs a main code file and a default Lakehouse.
- Eventstream: captures live events, transforms them without code, and routes them to an Eventhouse, a Lakehouse or an alert.
Decision rule: copy only, use a Copy job. No-code reshaping, use Dataflow Gen2. Code, use a Notebook. Many steps in order, use a Pipeline. Live events, use an Eventstream.
Match
Match each Northwind Rides task to the Fabric item for it
Pick an item on the left, then its match on the right.
Power BI on top, domains around it
Exam depth: recognize it.
At the end of the line sit Power BI items, which consume data rather than store it. A semantic model defines tables, relationships and measures (saved calculations). Reports are the charts built on top of it.
Direct Lake is a mode where a semantic model reads Delta tables straight from OneLake. There's no import copy. A refresh only updates metadata, the notes on which files hold each table, so it takes seconds.
A domain groups workspaces by business area, such as Finance or Operations. It helps people find data and lets admins delegate some settings. It doesn't grant access; workspace roles do that. Domains come back in a later lesson.
A workload is a set of related item types. The ones you use most are Data Factory, Data Engineering, Data Warehouse and Real-Time Intelligence.
Try it in Fabric
A free Fabric trial is enough for this.
- Sign in to the Fabric portal at app.fabric.microsoft.com.
- Select Workspaces, then + New workspace.
- Enter a unique Name, such as Northwind Rides Dev.
- Expand Advanced. Under Workspace type, select Fabric Trial. This assigns the workspace to your trial capacity.
- Select Apply. The new, empty workspace opens.
- Select + New item, then search for and select Lakehouse.
- Name it rides_lakehouse. Start with a letter, then use only letters, numbers and underscores. No spaces.
- Select Create. The Lakehouse opens with its Tables and Files folders.
Go back to the workspace list. You'll see two items: the Lakehouse and its SQL analytics endpoint, created for you.
Exam-style practice
You are a data engineer at Northwind Rides. A partner drops station maintenance CSV files in cloud storage every night. You need them in a Lakehouse each morning, unchanged, with the least setup. What should you use?
Pick one answer.
You are a data engineer at Northwind Rides. Trip data sits in large Delta tables in a Lakehouse. The ops team's Power BI report must show it without keeping a second copy of the data or waiting for long import refreshes. What should the semantic model use?
Pick one answer.
You are a data engineer at Northwind Rides. The finance team wants its own data area, separate from the ops team, with its own admins. A colleague suggests creating a second OneLake for finance. You need to meet the goal. What should you do?
Pick one answer.
Worth keeping
- One tenant has one OneLake, and every engine reads the same Delta tables in it.
- Workspaces hold items and run on a capacity; domains group workspaces.
- Lakehouse, Warehouse, Eventhouse and mirrored database store data.
- Pipeline, Copy job, Dataflow Gen2, Notebook, Spark job definition and Eventstream move or transform it.
- Power BI semantic models read lake data with Direct Lake, with no import copy.
On the job
Here's how the Northwind Rides platform might look after a few months.
A workspace called Rides data holds the engineering items. An Eventstream takes live trip events from the scooters and sends them to an Eventhouse. The ops team queries the last hour of trips there with KQL.
The customer and payment database in SQL Server is mirrored into Fabric. That gives a read-only copy that stays in sync, with no copy jobs to maintain. A nightly Pipeline runs a Copy job for the partner's maintenance CSVs, then a Notebook. The Notebook joins trips, stations and maintenance into clean Delta tables in a Lakehouse.
A second workspace, Ops reports, holds a semantic model in Direct Lake mode and the reports built on it. When the Notebook rewrites a table, the reports show the new numbers within seconds of a refresh.
Both workspaces sit in an Operations domain, so people can find them in the catalog. Nobody copied a table to make any of this work.
You now know the Lakehouse speaks Spark, the Warehouse speaks T-SQL, and the Eventhouse speaks KQL. Next lesson you ask all three the same question about Northwind's trips, and see how differently they say it.
Leaderboard
Loading…
50 XP for each lesson you finish, 10 for each right answer. Names show only for learners who turn on Show me on leaderboards.
Sources for this lesson (18)
Each fact in the lesson is marked with the number of the page it comes from.
- [4]Microsoft Learn: Study Guide for Exam DP-700
- [9]Microsoft Learn: What is Microsoft Fabric?
- [10]Microsoft Learn: Understand Microsoft Fabric licenses and capacity
- [21]Microsoft Learn: Fabric terminology
- [38]Microsoft Learn: What is OneLake?
- [39]Microsoft Learn: Workspaces in Microsoft Fabric and Power BI
- [40]Microsoft Learn: What is a lakehouse in Microsoft Fabric?
- [41]Microsoft Learn: Eventhouse overview
- [42]Microsoft Learn: What is Mirroring in Fabric?
- [43]Microsoft Learn: Choose an analytical data store in Microsoft Fabric
- [44]Microsoft Learn: What is Data Factory in Microsoft Fabric?
- [46]Microsoft Learn: Quickstart: Create a Copy job
- [47]Microsoft Learn: What is an Apache Spark job definition?
- [48]Microsoft Learn: Overview of Microsoft Fabric eventstreams
- [49]Microsoft Learn: Direct Lake overview
- [50]Microsoft Learn: Fabric domains
- [51]Microsoft Learn: Lakehouse tutorial: Create a Fabric workspace
- [52]Microsoft Learn: Create a lakehouse in Microsoft Fabric