Details
-
Improvement
-
Status: Resolved
-
Minor
-
Resolution: Fixed
-
2.7.2
-
None
-
Reviewed
Description
If you could make Hadoop Paths serializable, you can use them in Spark operations without having to convert them to and from URIs.
It's trivial for paths to support this; as well as the OS code we need to add a check that there's no null URI coming in over the wire, and test to validate round tripping
Attachments
Attachments
Issue Links
- is related to
-
HADOOP-13895 Make FileStatus Serializable
- Resolved
- links to