Dataverse tables enabling themselves on Link to Fabric

jose_henrique 

This is by default behavious, i checked with microsoft documentation and found this information.

By selecting Link to Microsoft Fabric, the system adds all nonsystem Dataverse tables that have the Track changes property enabled. You can add more tables later.

If you used this feature during public preview, you can continue to use the tables selected earlier. If you choose the Link to Fabric option in new environments (or unlink and relink existing environments), all tables are added.

Enabling this feature results in an increase in Dataverse database storage consumption. You can view additional storage consumption in the environment storage capacity details view in Power Platform admin center. For an example, you notice an additional file Account-Analytics if you selected the Account table for Fabric link. Also note that the chart only displays tables consuming the most storage. You can get a complete list of tables using the menu on the top right of the chart.

A practical workaround is to use two Lakehouses in Microsoft Fabric:

Create a Bronze Lakehouse to store all incoming data from Dataverse (even if it includes unwanted tables).

Create a second curated Lakehouse that contains only the tables required for your analysis and reporting.

This approach helps isolate unnecessary data while keeping your working environment clean and controlled.

If this helps, ✓ Mark as Kudos | Mark as Solution| Help Others

hey jose_henrique adding from other answers above. 

When you enable Link to Fabric:

1. Fabric creates shortcuts in OneLake
2. These shortcuts are driven by Dataverse metadata, not just your checkbox selection

Sometimes when Microsoft rolls out schema sync changes, new table classifications, system-table handling improvements and F&O schema updates, the Link to Fabric service will re-evaluate the source and rehydrate tables. Hence why you're seeing those additional 200 tables.

Since this is an expected behaviour from Microsoft Learn:
https://learn.microsoft.com/en-us/power-apps/maker/data-platform/azure-synapse-link-view-in-fabric

I suggest:

1. Treat Link to Fabric as “eventually consistent” (Expect occasional drift after service updates)
2. Separate ingestion from consumption, (+1 to Murtaza_Ghafoor)

   Bronze lakehouse - raw Link to Fabric shortcuts
   Silver lakehouse / Warehouse - only curated tables

   Avoid using the auto-linked tables for Power BI consumption and production workloads.

Appreciate if you can 'Kudos' and/or 'Accept as Solution' if this answered your query.