[
https://issues.apache.org/jira/browse/BEAM-5180?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16592727#comment-16592727
]
Ankur Goenka commented on BEAM-5180:
------------------------------------
That's correct. HDFS file URIs in this case do not have authority so // is not
required.
I agree with the resolution.
>From the documentation, the schema identification should be done only based on
>: but windows breaks this format.
We can have a separate way to identify windows file system by using
{code:java}
[a-zA-Z]:/{code}
given that windows can have only a single drive letter. We should also relax
back schema regex to just
{code:java}
(?<scheme>[a-zA-Z][-a-zA-Z0-9+.]*):.*
{code}
> Broken FileResultCoder via parseSchema change
> ---------------------------------------------
>
> Key: BEAM-5180
> URL: https://issues.apache.org/jira/browse/BEAM-5180
> Project: Beam
> Issue Type: Bug
> Components: sdk-java-core
> Affects Versions: 2.6.0
> Reporter: Jozef Vilcek
> Assignee: Ankur Goenka
> Priority: Blocker
> Fix For: 2.7.0
>
> Time Spent: 2h 40m
> Remaining Estimate: 0h
>
> Recently this commit
> [https://github.com/apache/beam/commit/3fff58c21f94415f3397e185377e36d3df662384]
> introduced more strict schema parsing which is breaking the contract between
> _FileResultCoder_ and _FileSystems.matchNewResource()_.
> Coder takes _ResourceId_ and serialize it via `_toString_` methods and then
> relies on filesystem being able to parse it back again. Having strict
> _scheme://_ breaks this at least for Hadoop filesystem which use _URI_ for
> _ResourceId_ and produce _toString()_ in form of `_hdfs:/some/path_`
> I guess the _ResourceIdCoder_ is suffering the same problem.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
