Hi Mang Zhang, Thanks for driving the proposal of Flink SQL support. This integration makes great significance to flink users for native vectorized execution. I have some questions about this proposal:
1. Could you provide overall architecture of flink integration in this proposal? There is no any introduction of Flink integration architecture. 2. Could describe proposed changes in more detail? The description of `Implementation` section is a little simple like `Support Flink RowData to Arrow`. 3. The interfaces like `SupportsAuronNative` listed in `Implementation` section are public interfaces for community developers. IMO, we could move these interfaces to `Public Interfaces`. WDYT? 4. The proposed changes of `Support Flink Built-in functions`, `Support Flink Connector` and `Support Native Format` are lack of change detail. For example, how to support flink connector? Which common interfaces should the developer implement? Looking forward to answers of above question. Regards, Nicholas Jiang On 2025/11/28 06:04:56 Mang Zhang wrote: > Hi, everyone > > > > > I'd like to start a discussion about AIP-1: Flink SQL + Auron Collaboration > [1]. > > Spark SQL on Auron has been highly successful, so I believe it's time to > explore Flink SQL on Auron. > > Looking forward to your feedback. > > > > > [1] > https://docs.google.com/document/d/1z3Noy4C9Lv54zEuL9MfGXohrDsIUh9zBpxOkmulRM6s/edit?usp=sharing > > > > > > > > -- > > Best regards, > Mang Zhang
