Github user jmahonin commented on the issue:
https://github.com/apache/phoenix/pull/423
@twdsilva Very impressive work, well done!
A few questions:
- Does this break compatibility with previous Spark versions? i.e., Spark
2.3.x wouldn't be binary-compatible with Phoenix 5.1.x (or whatever the next
version ends up being)
- It seems like much of the boiler-plate MR code has been replicated
yet-again into this new integration. The overall structure makes sense here,
but I'm wondering if some of that code shouldn't live in phoenix-spark at all?
---
