Fix sticky values in Dataflow Data Destination UI

Did you try opening the dataflow and click Save after syncing it from Git? Please vote for this Idea: https://community.fabric.microsoft.com/t5/Fabric-Ideas/Automatically-validate-Dataflow-Gen2-after-deployment/idi-p/4868304

frithjof_v, I'll give it a try next time. Interestingly, I've never had an issue updating the part of the data destination code that uses a workspace ID and lakehouse ID. I update these lines to refer to library variables using Variable.ValueOrDefault, and have confirmed they are working correctly in different workspaces without having to publish the dataflow after the sync.

Would you mind sharing how you're changing the table name through Git? What specific part of the mashup.pq document are you modifying?

 

Sharing any repro steps will help us get to the bottom of it.

For example, like this: shared DimTABCRoster_DataDestination = let Pattern = Lakehouse.Contents([CreateNavigationProperties = false, EnableFolding = false]), Navigation_1 = Pattern{[workspaceId = workspace_id]}[Data], Navigation_2 = Navigation_1{[lakehouseId = renamed_lakehouse_id]}[Data], TableNavigation = Navigation_2{[Id = "DimTABCRoster", ItemKind = "Table"]}?[Data]? in TableNavigation; I was editing the Table Navigation step in the Git repository, where it filters to the table as "DimTABCRoster". Now that I take a closer look at it, it seems this step is referencing an existing table rather than creating a table. I would personally like for this step to create a new table if none exists with that name. It is more similar to the UI experience where a new table is created the first time you run the dataflow with a newly named destination.

It is possible to change the table name in Git. However, we woudl recommend making the changes through the new Advanced Editor for the data destination query so you don't have to manipulate the whoel document manually or require Git:

Advanced edit for data destination queries in Dataflow Gen2 - Microsoft Fabric | Microsoft Learn

 

In the TableNavigation step we actually use the question marks (?) as a way to throw a null if the table isn't found. Depending on how the rest of your M Script is set, it may actually create a new table or not.

 

The article below goes a bit more in-depth about the Settings that the data destination query could have and how such settings are set as an attribute record to the query that originates the data:

Dataflow Gen2 data destinations and managed settings - Microsoft Fabric | Microsoft Learn

 

For more questions about the behavior of data destinations, we recommend posting the questions in the Data Factory forum.