Hi together,
we spent a lot of time investigating the use of Fabric and I really enjoyed using it from a functional point of view, having everything in one place.
But doing some bigger workloads just dropped my excitement a little bit. How is your experience on this one? We made a DWH on Fabric with a couple tables (derived from the TPC H benchmark) with 1 GB of volume (biggest table 1,500,000 rows). Ingestion with Dataflow was faster than Copy activity (didn't expected that), but query time with the Fabric DWH on F32 CU (5-6 $ per hour) was 3-4 times longer than on Snowflake (with Compute 4$ per hour).
It's expensive + slow. Somehow the only selling point is PowerBI and Dataflow I feel. Would like to hear your opinions and findings.