Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
-
None
Description
The current spark interpreter is pretty complex. Because of 2 reasons:
- It is written by java and it uses a lot of java reflection to call spark’s scala api.
- It mix both scala 2.10 & scala 2.11.
With more spark versions are released and more scala version supported (spark 2.3 will support scala 2.12). Zeppelin’s spark interpreter would become more and more complex and pretty hard to maintain. So I propose to refactor the spark interpreter.
Attach the design doc. \cc moon felixcheung jongyoul
Attachments
Issue Links
- blocks
-
ZEPPELIN-2385 Release 0.8.0
- Resolved
- is related to
-
ZEPPELIN-3807 Fix spark interpreter on Windows
- Open
- links to