Hello,

I need to set up some buckets in GCP which will be used as an analytics and productive data lake. I am getting diverging feedback on whether hierarchical namespaces (HNS) should be enabled for these buckets.

On one hand, HNS is advisable for analytics, partitioned table formats like Delta and Iceberg, etc. Google-employed experts are advising it.

On the other, other other people say they know from Databricks that HNS is not supported. There are also comments that it does not work with Unity Catalog on GCP, and UC is a feature I will need to have.

Can someone please give me a definitive answer of what works and when with Databricks on GCP concerning HNS?