I have a #C03JKTFFX0S question 🙂 When it comes to open table formats (OTF), do folk tend to see Flink used with one format more often than another? Or is it Iceberg/Hudi/Delta Lake equally shared?

First of all it's a pity seeing no mention of apache paimon in the Flink slack, when paimon was born directly from the Flink umbrella to be the best connector for Flink. But to address your question (this is on view and experience, others might interact with different company/customer profiles): Pesonally I'm only aware of two production deployments of Hudi (in the EU/US - although i know there are more), but overall it doesn't pop up in customer discussions (I have participated in at least 50+ Flink opportunities). There were many Hudi installations in China, but many of them have migrated to paimon (due to the scale in china, iceberg can't meet the needs there and is only used in medium to small size companies). Delta might come up every now and then only if people are already using spark and want to keep some consistency, but overall i would say the flink-delta adoption is close to zero. So for the most part it's Iceberg, but with the recent adoption of paimon that might change soon (speaking strictly for the Flink ecosystem).