Calcite and STREAM SQL support

[akennedy4155]

I came across Apache Calcite and I'm wondering if this is a standard that we would want to implement for Arroyo. It has a great library and documentation for different types of functions and aggregations for streams, and I think, based on my research, that Flink integrates and supports the functionality defined by Calcite.

I don't know the extent of the implementation of Calcite. I know that it's a lot of Java, but I'm not sure if the operations are actually implemented or if it's mostly parsing, planning, and structure, like I understand DataFusion to be, or if it actually implements these functions.

Either way, I'd be curious to hear what devs think about this.

---

Please correct me if there's any misinformation about DataFusion or Calcite embedded in the paragraph above. I'm still learning about this space and could be understanding something wrong.

[akennedy4155]

Just found this in DataFusion docs.... So I think I'm definitely misunderstanding.

In that respect, DataFusion is inspired by Apache Calcite in the Java world.

[akennedy4155]

One difference that I can find between the two is that Calcite is more complete than DataFusion with streaming aggregations and functions. Aggregations and windows such as TUMBLE, HOP, Sliding Windows, Cascading Windows, Joins etc. are defined in Calcite but not present in DF.