According to the docs, https://learn.microsoft.com/en-us/power-bi/enterprise/directlake-overview#create-a-delta-table-in-lakehouse, a so-called "VORDER" Compression is needed or very much recommended to use Direct Lake Feature. Since this is on the storage format and the whole point of a data lake is to have the storage in an open format, Microsoft should really open source this compession. Open source means both Spec and a reference implementation, as Databricks does with the Delta Lake Format itself.
I do want to have different engines that understand VOrder
2 Comments
Another option might be licensing the algorithm to other vendors that plan to integrate into Fabric (e.g. Databricks, ESRI, Neo4j). Opensourcing or licensing: The most important would be for customers to be able to use it seamlessly to truly leverage Direct Lake.
- Status added:New
