Optimizing storage with shortcuts and mirroring
1. Optimizing storage with shortcuts and mirroring
Welcome back! So far, you have copied data into Fabric with a Copy job and a pipeline. Next, you'll see two alternatives: shortcuts, which reference data without copying it, and mirroring, which keeps a synchronized replica of a source.2. Every shortcut has two paths
Every shortcut has two paths: where users see it, and where the data actually lives. Consider an analytics lakehouse called LH_Analytics that exposes curated product data through a shortcut. The shortcut path is the fully qualified path where users see that reference, so it contains LH_Analytics. The target path is the fully qualified path of the item that stores the original data. If the original data is stored in LH_Curated, that name appears in the target path. Products is the local shortcut name. It can differ from the source table name, which is why the target path ends with DimProduct even though the shortcut is called Products.3. Internal OneLake shortcuts
An internal OneLake shortcut references data in another Fabric item. Both sides are inside the Fabric OneLake environment. For example, an analytics lakehouse can expose curated data that remains in a source lakehouse, or a warehouse in another workspace. To set one up, open the lakehouse where the shortcut should appear, go to Files, and select New shortcut to open the New shortcut window. Choose the internal Microsoft OneLake option and keep the default Passthrough identity, so each user's own access to the source applies. If the source lakehouse was created with the Lakehouse schemas option, its tables sit under a dbo schema that you can select in one go. The reference appears in the analytics lakehouse while the original data remains in the curated source.4. External shortcuts
An external shortcut is a secure connection from Fabric to supported storage outside OneLake. The data may live in another cloud storage service rather than in a Fabric item. Fabric needs the correct connection and authentication details to reach it. The result is still a reference, not a copied dataset. External shortcuts support many sources, such as Amazon S3.5. Shortcut permissions
Creating a shortcut doesn't grant access to the data it points to. A user needs permission to create the shortcut in the destination item. The user also needs an authorized way to read the target data. With Passthrough, Fabric checks the user's target permission each time data is read, so if that access is removed later, the shortcut stays listed but queries fail. External shortcuts use the connection's credentials instead. The key rule is that creating a shortcut and reading the data behind it are separate permissions.6. Reference or replicate external data
Here's how a shortcut differs from mirroring. A shortcut is a reference to the original data. Fabric creates no separate copy, so the original location stays authoritative and reads reflect its latest data. Mirroring is different. It maintains a replicated copy from a supported operational source, such as a SQL database. Fabric keeps that replica in sync as changes arrive from the source.7. Choosing the right option
The choice comes down to where the data lives and whether a local copy is required. Use an internal shortcut for data already stored in OneLake. Use an external shortcut for supported storage outside OneLake. Use mirroring for a supported operational database, which a shortcut can't reference, so Fabric keeps a replicated copy in sync.8. Let's practice!
Now, let's practice!Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.