modeling streaming and other extended relational features

[zhouqingqing]

steaming: unbounded data, unbounded processing

• trigger, window and watermark
• grouped window functions (fixed|tumble, slide|hop, session)
• grouped auxiliary functions (_start|end)
• MATCH_RECOGNIZE

unbounded processing:

• when to output
• optimizer: each data snapshot has a plan, use which plan?

batch: bounded stream and delta maintanence

---

Progress: implements fixed and slide window

[zhouqingqing]

Json query:

• path.subpath.subsubpath = 'something'

object method:

• select strcol.substring(1,3) from ... good for intellisense

Relational Matrix Algebra: https://arxiv.org/abs/2004.05517