Fabric - Warehouse or Lakehouse
Thomas LeBlanc
Microsoft Fabric Architect | Business Intelligence Architect | Microsoft Data Platform MVP | Power BI Super User | Speaker | Mentor | Technical Business Strategist | Author
During a webcast by Kasper on BI, Bogdan Vricat (VP @ MSFT), talks about many things with Microsoft Fabric. There are no demos, just an hour+ of good chat of when, why, where and how to use Fabric.
One part that really caught my attention was the difference between a Warehouse and a Lakehouse. Basically, the Lakehouse is for the python data warehouse developer whereas the Warehouse is for the T-SQL data warehouse developer.
An interesting conversation in this recording is the reasoning for Fabric. It is a place for bringing all types of data warehouse developers into one system - OneLake.
The mention of OneLake like OneDrive is shown by using the add-in for Windows Explorer to show the files (parquet and Delta Table with the transaction log).
Even though you do not need to use that for copying files or modifying files, it is nice to see the behind the scenes of the delta tables.
In the above image, there is a Warehouse (left) and Lakehouse(right). The Warehouse looks more like a SQL Server Management Studio view of a database and schema drilling into tables or views. The Lakehouse view is more like a Synapse view of a connection to an Azure Data Lake Storage Gen2 where you can query the parquet files or look at Delta tables that are created from the parquet files.
Both have a means to create pipelines as well as creating a Power BI dataset. It just helps with those who are familiar with a database versus data lake.