Hi, the documentation is on GitHub, but be careful which repo. The original microsoft/bc2adls is read-only now. The one that's still maintained is the fork from Bert Verbeek (Bertverbeek4PS/bc2adls), and it also supports Fabric lakehouse. The setup guide is in there.
Setup is not hard. You make an app registration and a storage account, then in BC you pick the tables and fields, and schedule the export with a job queue.
The delta part works on the system fields, so only records changed since last export are sent. Deletes are tracked separately. On the Azure side a Synapse pipeline or a Fabric notebook merges the deltas into the final tables.
Two tips. The first full export of big tables like G/L Entry or Value Entry can take long, so run it out of hours. And export only the fields you really need.
For your second question, no, it's one way only. Nothing supported writes from the lake back into BC. For that you use the APIs, standard or custom API pages, called from Data Factory or Power Automate. A restore from the lake I wouldn't try at all.
Glad to help - follow up if anything is unclear.
► If this solved it, marking it verified helps others too.
Regards,
Grigorios Mavrogeorgis
Business Central Consultant & AL Developer
Work: Gmsoft Limited
Blog: insidebusinesscentral
LinkedIn: linkedin.com/in/gregorymavrogeorgis
