JingsongLi commented on code in PR #9608:
URL: https://github.com/apache/paimon/pull/9608#discussion_r3986881659
##########
paimon-python/pypaimon/filesystem/resolving_file_io.py:
##########
@@ -137,7 +138,18 @@ def write_row(self, path: str, data, fields=None,
zstd_level: int = 1, **kwargs)
return self._get_fileio(path).write_row(path, data, fields,
zstd_level, **kwargs)
+ @property
+ def uri_reader_factory(self):
Review Comment:
Java's FileIO does not own a UriReaderFactory; the factory wraps FileIO.
Could we keep that ownership direction here? The new read paths already call
`UriReaderFactory.from_file_io(file_io)` directly, so this property is not
required for the read fix.
I verified that this addition fixes an existing write-path assumption:
`BlobFileWriter._to_blob()` accesses `self.file_io.uri_reader_factory`, which
raises `AttributeError` for ResolvingFileIO on the base version. That is a
separate write-path compatibility issue. Could we handle it separately and pass
the factory explicitly at the BLOB reader/writer boundary? This would avoid
extending FileIO with factory state and lifecycle handling as part of this
read-path change.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
