Direct lake semantic models do not automatically reflect schema changes in the underlying lakehouses. (Link)
Metadata can be refreshed by editing the model tables, refreshing the table list, and then saving. Opening and editing all semantic models downstream of a lakehouse every time a change occurs is not a scalable solution.
Please provide a simple way to programmatically refresh this metadata in Semantic Models via a REST API call or similar.
2 Comments
- Hello Sean: I believe you can accomplish what you want using semantic link. https://learn.microsoft.com/en-us/fabric/data-science/semantic-link-overview
- todd-wilson - unfortunately, I believe sempy's refresh function only kicks off a framing operation in a direct lake model (https://learn.microsoft.com/en-us/fabric/fundamentals/direct-lake-how-it-works#framing) - the table metadata must be refreshed manually by opening/editing the direct lake model table list.
