Github user kaspersorensen commented on the pull request:
https://github.com/apache/metamodel/pull/79#issuecomment-164188785
You're right about FTP, normal files and S3. We could add those schemes
too, if that's what you mean? But the way of accessing them will be through
HDFS - or at least via the _hadoop (pause) FileSystem_ (class). If it makes
things more clear, we could create a new class instead, called "HadoopResource"
which would cover all the different schemes (and could then hold _just_ an URI
instead of the individual URI elements).
I just now tried to see if I could do the URI assembly in a more reusing
way. Encountered the following small issues:
* UriBuilder is part of JAX-RS (and there is also one in apache HTTP
components). I think it would be unfortunate to depend on an optional library
just to do this.
* So then I tried instead the URI class...
* That requires me to catching a URISyntaxException which will never
happen. A bit annoying IMO.
* It changes a small thing in one of the test cases. In the case of
```emrfs:///home/metamodel.txt``` it shortens it by removing two slashes,
making it just ```emrfs:/home/metamodel.txt```. I guess that's a valid change
from a pure URI point of view, but it's not what I have ever seen in the
documentation/examples for these kinds of schemes.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
