All 35 lessons
Start here: cloud, data, Fabric and Power BI 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. Power BI in plain words: semantic models, reports and DAX
Prepare data: get data in (exam area 2, 45-50%)
- 6. Connect to data and find it: connections, gateways, OneLake catalog and Real-Time hub
- 7. Choosing a data store: lakehouse, warehouse or eventhouse
- 8. Use data without copying: shortcuts, mirroring and OneLake integration
- 9. Ingest data: Dataflow Gen2, pipelines, Copy job and notebooks
- 10. Power Query in Dataflow Gen2: filter, data types, merge and new columns
- 11. Messy data: duplicates, missing values and nulls
Prepare data: transform and query (exam area 2)
Design and build semantic models (exam area 3, 25-30%)
- 18. Choose a storage mode: Import, DirectQuery, Direct Lake and Dual
- 19. Relationships: star schemas, bridge tables and many-to-many
- 20. DAX foundations: measures, filter context, CALCULATE and variables
- 21. DAX functions the exam tests: iterators, table filters, windows and information functions
- 22. Calculation groups, dynamic format strings and field parameters
- 23. Composite models and large semantic model storage format
Optimize models, then secure them (exam areas 3 and 1)
- 24. Direct Lake in depth: fallback, refresh and the two flavors
- 25. Incremental refresh for semantic models
- 26. Faster reports and DAX: finding and fixing slow queries
- 27. Who can see what: workspace roles and item permissions
- 28. Fine-grained security: rows, columns, objects and files
- 29. Governance: sensitivity labels and endorsement
Lifecycle, then exam practice (exam area 1)
20 min
Use data without copying: shortcuts, mirroring and OneLake integration
Access data where it lives with shortcuts and mirroring, and make Eventhouse and semantic model data available in OneLake.
By the end of this lesson you can
- Create an internal or external shortcut in the right Lakehouse folder
- Explain what mirroring copies, how fresh it is and what it costs
- Turn on OneLake availability for a KQL database and OneLake integration for a semantic model
- Choose between a copy, a shortcut, mirroring and OneLake integration
From earlier lessons · 1 of 2
You run an INSERT through a Lakehouse's SQL analytics endpoint. What happens?
Pick one answer.
From earlier lessons · 2 of 2
Juniper's finance team writes T-SQL and needs two tables to update together or not at all. Which store fits?
Pick one answer.
Juniper Coffee's coffee-bean supplier keeps its price lists in its own Amazon S3 bucket. The till database in Azure SQL must feed reports without slowing the tills. The loyalty events sit in an Eventhouse, and finance's budget model holds tables cleaned in Power Query. Your manager asks: "Can we use all of it without building a single copy job?"
Shortcuts: a pointer, not a copy
Exam depth: know it cold.
A shortcut is an object in OneLake that points to data stored somewhere else. Nothing is copied. It looks like a folder, so Spark, the SQL analytics endpoint, KQL and Power BI Direct Lake can all read it. You create shortcuts in a Lakehouse or a KQL database. Deleting a shortcut removes only the pointer; renaming or deleting the target breaks it.
| Internal shortcut | External shortcut | |
|---|---|---|
| Points to | Lakehouse, Warehouse, KQL database, mirrored database, SQL database, semantic model. | ADLS Gen2, Azure Blob Storage, Amazon S3, S3 compatible, Google Cloud Storage, Dataverse, OneDrive and SharePoint, Iceberg. |
| Signs in with | Your own identity by default. | A saved cloud connection, from lesson 6. |
The item types don't need to match: a Lakehouse shortcut can point to a Warehouse table, in any workspace.
Shortcut caching (recognize it) keeps files read from S3, S3 compatible, Google Cloud Storage and gateway shortcuts in OneLake for 1 to 28 days. That cuts cross-cloud read fees. Turn it on in Workspace settings > OneLake.
You are an analytics engineer at Juniper Coffee. The bean supplier's price-list files must stay in its Amazon S3 bucket, by contract. You need them visible in your Lakehouse with no copy. What should you create?
Pick one answer.
Exam depth: know it cold.
A Lakehouse has two top folders, and shortcuts behave differently in each.
| In Tables | In Files | |
|---|---|---|
| Where you create it | Top level only. | Any folder level. |
| What it points to | Data in a table format, such as Delta. | Data in any format. |
| Becomes a table? | Yes, if the target is Delta. | No. Spark reads it by path. |
| SQL analytics endpoint and Direct Lake see it? | Yes. | No. |
Rule: point a Tables shortcut at Delta data you'll query with SQL or Power BI. Point a Files shortcut at raw files, such as CSV, that Spark will process.
Exam depth: recognize it.
Through a shortcut you get the most restrictive of two sets of permissions: on the shortcut's location and on the target. By default an internal shortcut passes your identity to the target (passthrough). An external shortcut always uses a connection's fixed credential (delegated). You can also pick Delegated identity for an internal shortcut when you create it.
One exception matters for Power BI. A semantic model in Direct Lake on SQL mode reads internal shortcuts with the item owner's identity, not the viewer's. Direct Lake on OneLake passes the viewer's identity. Lesson 24 compares the two.
Mirroring: a live, read-only copy of a database
Exam depth: know it cold.
Mirroring continuously replicates an outside database into OneLake as Delta tables. You pick the whole database or chosen tables, and Fabric keeps them in sync. You build no ETL pipeline. (ETL means extract, transform, load: the usual copy-and-reshape jobs.) Changes can arrive as often as every 15 seconds, so it's near real time.
You get a mirrored database item plus a SQL analytics endpoint. The mirrored tables are read-only: read them with SQL, Spark or Power BI Direct Lake, but don't edit them.
Sources include Azure SQL Database, Azure SQL Managed Instance, SQL Server, Azure Cosmos DB, Azure Database for PostgreSQL, Snowflake, Oracle, Google BigQuery and SAP.
Cost: replication compute is free. Storage is free up to 1 TB per capacity unit (an F64 gets 64 TB). Queries cost capacity as normal. The capacity must be running, or nothing replicates.
Recognize it: metadata mirroring copies only the catalog and reads data through shortcuts (Azure Databricks). Open mirroring lets your own app write change files into a landing zone.
You are an analytics engineer at Juniper Coffee. Store managers need till sales from the Azure SQL Database in reports within a minute of each sale. Reports only read the data. You must not slow the tills, and you have no budget for ETL jobs. What should you use?
Pick one answer.
OneLake availability for an Eventhouse
Exam depth: know it cold.
By default, a KQL database's data isn't in OneLake as Delta, so Spark, T-SQL and Direct Lake can't read it there. OneLake availability is the opt-in switch. It writes a logical copy of the data to OneLake in Delta format, at no extra storage cost.
You turn it on for a whole KQL database or one table. Select it, then in the OneLake section of the details pane set Availability to Enabled. At database level, new tables are included, and Apply to existing tables backfills old ones.
Once it's on, you can:
- Create a shortcut to the tables from a Lakehouse or a Warehouse.
- Build a Power BI model on them in Direct Lake mode.
- Read them from a notebook, or from a SQL analytics endpoint under Analyze data with.
The Delta copy is read-only. The KQL database's retention policy also applies to it.
You are an analytics engineer at Juniper Coffee. Loyalty events land in a KQL database. You need them in a Direct Lake semantic model next to Lakehouse sales tables. Data must not be copied by a pipeline, and a few hours' delay is fine. What should you do?
Pick one answer.
OneLake integration for semantic models
Exam depth: know it cold.
An Import semantic model loads a copy of its data into Power BI's memory. Often that data was cleaned in Power Query and exists nowhere else. OneLake integration writes the model's import tables to Delta tables in OneLake, so Spark, T-SQL and other tools can reuse them.
- Needs: a workspace on a Power BI Premium P or Fabric F capacity. Not Pro, Premium Per User or Embedded.
- Turn on: in the semantic model settings, expand OneLake integration, set it On and select Apply.
- Then refresh. Nothing is written until a manual or scheduled refresh runs. Each later refresh updates the Delta tables.
- Use it: in a Lakehouse, open the ... menu on Tables, select New shortcut > Microsoft OneLake, pick the semantic model and its tables.
The files sit in a folder named after the model, ending in .SemanticModel. Creating those shortcuts needs Contributor-level (read, write, explore) permission on the model.
You are an analytics engineer at Juniper Coffee. Finance's import semantic model, on an F64 capacity, holds budget tables cleaned in Power Query. Data scientists need those exact tables in a notebook, and nobody should rebuild the cleaning logic. You turned on OneLake integration. What else must happen?
Pick one answer.
Ingest or access: the decision table
Exam depth: know it cold.
Ingest means you copy data into a Fabric store. Access means you read it where it already lives.
| Copy (pipeline, Copy job, dataflow) | Shortcut | Mirroring | OneLake availability / integration | |
|---|---|---|---|---|
| Data moves? | Yes. | No. | Yes, a read-only replica. | A Delta copy written for you. |
| Source | Almost anything. | Files or Delta in storage or OneLake items. | An outside operational database. | A KQL database / an import semantic model. |
| Freshness | As of the last run. | Always current. | Near real time. | Minutes to hours / last model refresh. |
| Reshape on the way? | Yes. | No. | No. | No. |
| Writable target? | Yes. | Only if the target allows. | No. | No. |
Decision rule: "no data movement" means a shortcut. A live, read-only copy of an outside database means mirroring. KQL or model data needed in OneLake means OneLake availability or integration. A schedule, transforms or a writable table means a copy.
Try it in Fabric
A free Fabric trial is enough.
Create a shortcut
- Create a Lakehouse named juniper_sales: New item > Lakehouse > Create.
- Download the sample file fabrictutorialdata.blob.core.windows.net/sampledata/Coffee/Dim_Products.csv.
- In the Lakehouse, select Get data > Upload files, and upload it.
- In Files, open the file's ... menu, select Load to Tables > New table, name it dim_products and select Load.
- Create a second Lakehouse named juniper_marketing.
- Open the ... menu on its Tables folder and select New shortcut.
- Select Microsoft OneLake, pick juniper_sales, select Next. Keep Pass-through if asked for a connection method.
- Expand Tables, select dim_products, select Next, then Create.
The table appears with a link icon, and no data was copied.
Turn on OneLake availability (needs an Eventhouse)
- Open a KQL database and select it, or one of its tables.
- In the OneLake section of the details pane, set Availability to Enabled.
- In Enable OneLake availability, select Enable.
Exam-style practice
You are an analytics engineer at Juniper Coffee. OneLake availability is on for the loyalty KQL database. You now need to rename the ScanEvents table, and the rename command fails. What should you do?
Pick one answer.
You are an analytics engineer at Juniper Coffee. The marketing Lakehouse needs the finance workspace's store Delta table. A Direct Lake semantic model in marketing must read it, and the data must not be duplicated. What should you create?
Pick one answer.
Worth keeping
- A shortcut points to data without copying it; Tables shortcuts to Delta become queryable tables.
- Mirroring keeps a read-only, near real-time Delta copy of an outside database, with free replication.
- OneLake availability puts KQL data in OneLake as Delta, with a delay of up to 3 hours by default.
- OneLake integration writes an import model's tables to Delta on each refresh, on P or F capacity.
- Read the constraint word: no data movement, near real time, or reuse the model's data.
Some data can't be reached in place. The supplier emails a new price list every week, and it must land in a table on schedule. Next lesson: Dataflow Gen2, pipelines, Copy job and notebooks, and how to pick between them.
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 (10)
Each fact in the lesson is marked with the number of the page it comes from.
- [6]Microsoft Learn: Study guide for Exam DP-600: Implementing Analytics Solutions Using Microsoft Fabric
- [40]Microsoft Learn: What is Mirroring in Fabric?
- [81]Microsoft Learn: Microsoft Fabric decision guide: choose a data store
- [83]Microsoft Learn: OneLake shortcuts
- [84]Microsoft Learn: OneLake shortcut security
- [85]Microsoft Learn: Create an internal OneLake shortcut
- [86]Microsoft Learn: Quickstart: Get data into OneLake
- [87]Microsoft Learn: Microsoft Fabric decision guide: Choose a data movement strategy
- [88]Microsoft Learn: Turn on OneLake availability for an eventhouse
- [89]Microsoft Learn: OneLake integration for semantic models