patricklucas commented on issue #11990:
URL: https://github.com/apache/iceberg/issues/11990#issuecomment-3004231324
I also ran into this. I created a namespace like `Namespace.of("foo",
"bar.baz")` expecting that the namespace "levels" would be treated opaquely
like with other catalogs. However, after a round trip to the `iceberg_tables`
table, this namespace is now interpreted as `Namespace.of("foo", "bar", "baz")`.
In the `RESTCatalog`, for instance, when namespace are included in URLs,
each level gets URL-encoded and then the levels are joined by `%2f` (i.e. `/`),
so dots would be preserved. That said, I'm not sure what the behavior is if if
a namespace level includes a slash.
Nevertheless, Iceberg should take a clear stance on namespace handling that
is consistent across catalog implementations. The most explicit approach would
be to say that namespace levels may contain any Unicode character, and then
catalog implementations always act accordingly, encoding and decoding as
necessary. This would require each implementation to choose how to represent
them, and "double encoding" might be required, for instance to solve the
potential slashes issue in `RESTCatalog`.
Example: encoding namespace `Namespace.of("foo/bar", "baz")` for inclusion
in a URL, becomes `foo%252Fbar%2Fbaz`:
```
In [1]: import urllib
In [2]: level1 = urllib.parse.quote_plus('foo/bar')
In [3]: level1
Out[3]: 'foo%2Fbar'
In [4]: level2 = urllib.parse.quote_plus('baz')
In [5]: level2
Out[5]: 'baz'
In [6]: encoded = urllib.parse.quote_plus(level1 + '/' + level2)
In [7]: encoded
Out[7]: 'foo%252Fbar%2Fbaz'
In [8]: [urllib.parse.unquote(level) for level in
urllib.parse.unquote(encoded).split('/')]
Out[8]: ['foo/bar', 'baz']
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org
For additional commands, e-mail: issues-h...@iceberg.apache.org
