Fabric direct lake data models development with local git repo

[morz3d]

Hi all,

I am curious about best practices out-there.

I am evaluating Fabric capacities now, and i want to explore possibilities to develope my semantic models as well as keep changes in our local git repo.

So far I see only 2 options to interact with model in terms of development.

• having DEV semantic model created in the service and developed from the service. I am on purpose avoiding write to this model with XMLA endpoints so it is editable from workspace. Once changes are done i use Tabular Editor to read this model and sync with local git repot folder structure. To promote changes to UAT and PROD workspaces I use ALM toolkit so i have more control there.
• Different approach would be to use tabular editor to create DEV semantig model, so i loose ability to edit model from workspace, while promote to UAT and PROD would remaint the same.

Are there alredy some golden stadards here, can someone share their tought on this topic, or their current setup. 

[Anonymous]

Hi @morz3d 

For teams prioritizing version control and detailed change tracking, starting with Tabular Editor and leveraging git for source control from the outset is beneficial. For environments where rapid prototyping and ease of access are key, developing in the DEV workspace with subsequent synchronization to git might be preferred.

Best Regards,

Jayleny

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[Anonymous · Accepted solution]

Until the dust settles with Fabric and Git integration, you will have to pick the method you can support. I think the 1st option above accomplishes the 'Edit in service' requirement, but does place extra work in the capturing of artifacts to a Repo. The 2nd options makes the work less complicated but does not enable the 'Edit in service' requirement