Understanding Lineage View

[mattbenj]

I have multiple Dataflows that read data from the same Lakehouse and write to the same Warehouse. Why do the Warehouse and Lakehouse show up as different items in Lineage view (second photo)? Why don't they show as 1 node line my first screenshot?

[mattbenj]

@v-karpurapud I think there is a bit of a misunderstanding. I don't expect a Lakehouse and Warehouse to appear as one node. In this screenshot you see 3 Dataflow Gen2s. Each one reads data from a different table from the same Lakehouse (StagingLakehouse) amd writes data to three different tables in the same Warehouse (fabric_finance). I was trying to understand if there is a way to get the lineage view to show something like this. 

 

This works like I expected for some of my other Dataflows that write to the Lakehouse. 

[v-karpurapud]

Hi @mattbenj 

Thank you for clarifying. The scenario you described, where a single Lakehouse feeds multiple Dataflow Gen2s that merge into one Warehouse node, is a valid expectation, and lineage will show this when multiple dataflows write to a Lakehouse.

However, for Warehouse destinations, Fabric lineage treats each Dataflow Gen2 load as a separate event. Even if all dataflows read from the same Lakehouse and write to different tables in one Warehouse, the lineage view displays separate paths instead of combining them into a single Warehouse node.

This is the current expected behavior and reflects how Warehouse loads are modeled in Fabric lineage, not differences in connection or configuration. Currently, there is no supported way to visually combine multiple Dataflow Gen2 to Warehouse paths into one Warehouse node.

Let us know if you require any additional assistance or clarification.

Regards,

Microsoft Fabric Community Support Team.

[v-karpurapud]

Hi @mattbenj 

We have not received a response from you regarding the query and were following up to check if you had the opportunity to review the information provided. Please feel free to contact us if you have any further questions.

Thank You.

[v-karpurapud]

Hi @mattbenj 

I wanted to check if you’ve had a chance to review the information provided. If you have any further questions, please let us know. Has your issue been resolved? If not, please share more details so we can assist you further.

Thank You.

[GauravLZ888]

Hi @mattbenj 

Per my understanding, In Fabric, the Lineage view merges items into a single node only when it is 100% certain they represent the exact same logical entity.
It would otherwise splits nodes when there is a mismatch in how the connections are defined.

therefore the possible cause of this issue is Connection Mismatch. I am guessing here Your nodes are likely split because of either of the following reasons:

1. Different Connectors used in dataflows. e.g. One Dataflow might use the native Fabric Lakehouse connector, while another uses the SQL Connection String.

2. Staging Differences: If "Enable Staging" is turned on for some Dataflows but off for others, Fabric treats the data paths as distinct architectures.

3. Auth Methods: Using an Organizational Account in one and a Service Principal in another can cause a split.

possible option to fix it: i.e.to get the single-node look from your first screenshot:

1. Standardize the Destination: Open your Dataflows and ensure the Data Destination is selected via the Navigator (choosing the workspace/item from the list) rather than pasting a URL.

2. Match Staging Settings: Ensure all Dataflows writing to that Warehouse have the same "Enable Staging" toggle setting.

Hope this helps!

[v-karpurapud · Accepted solution]

Hi @mattbenj 
Thank you for contacting the Microsoft Fabric community forum.
 

Fabric lineage is shown at the artifact level, so Lakehouse and Warehouse appear as separate nodes even when they share the same data in OneLake. When Dataflow Gen2 reads from a Lakehouse and writes to a Warehouse, lineage represents this as two distinct steps to reflect data movement between different item types.
Here is the reference:Lineage in Fabric - Microsoft Fabric | Microsoft Learn
 

A Lakehouse also includes a SQL analytics endpoint, which can appear as an additional node and sometimes cause confusion. Multiple nodes for what appears to be the same source or destination typically occur when different connection methods are used (such as Lakehouse connectors, SQL endpoints, or Warehouse connections), when lineage is viewed at different levels, or when similarly named items exist in different workspaces.
Better Together - the Lakehouse and Warehouse - Microsoft Fabric | Microsoft Learn
 

Currently, there is no supported way to merge Lakehouse and Warehouse into a single node in lineage. To maintain a clear and consistent lineage view, all dataflows should use a consistent connection approach and reference the exact same Lakehouse and Warehouse artifacts.

If you have any further questions, feel free to reach out and we’ll be glad to assist. If I’ve misunderstood any part of your scenario, please let me know.

Regards,

Microsoft Fabric Community Support Team.