[Python] Read ORC metadata

[Fokko]

Describe the enhancement requested

When reading an ORC schema, the metadata of the fields isn't exposed. For more information check: apache/iceberg#6973 (comment)

Component(s)

Python

[jorisvandenbossche]

In C++, there is a ORCFileReader::ReadMetadata(), which is exposed as the metadata property in the Python ORCReader. Does that already give the relevant information, or is this a different set of metadata? (I know that arrow and parquet also both have file-level metadata as field-level metadata, so the same might be the case here)

I see that the ORC C++ orc::Type has a getAttributeKeys, and the proto spec has an attributes field for a Type (https://github.com/apache/orc/blob/cd79029e41e9cfe434529f820e73900e6a72797e/proto/orc_proto.proto#L226). Is this the information that you looking for?

[wgtmac]

In C++, there is a ORCFileReader::ReadMetadata(), which is exposed as the metadata property in the Python ORCReader. Does that already give the relevant information, or is this a different set of metadata? (I know that arrow and parquet also both have file-level metadata as field-level metadata, so the same might be the case here)

I see that the ORC C++ orc::Type has a getAttributeKeys, and the proto spec has an attributes field for a Type (https://github.com/apache/orc/blob/cd79029e41e9cfe434529f820e73900e6a72797e/proto/orc_proto.proto#L226). Is this the information that you looking for?

I think @Fokko is looking for a way to expose attributes from ORC. attributes store the information about column field-id defined by Apache Iceberg.