thinkharderdev commented on issue #5770: URL: https://github.com/apache/arrow-rs/issues/5770#issuecomment-2118124262
> > I don't think so because you need the column chunk metadata to actually read the column. > > https://github.com/apache/parquet-format/blob/master/src/main/thrift/parquet.thrift#L863-L867 `ColumnChunk` is neccessary, but `ColumnChunkMeta` is `optional`. This might making io-estimating a bit tricky but it making `ColumnChunkMeta` not in footer? And you're right because we need the column chunk metadata to actually read the column. Yeah, you're right and I think I just misunderstood the original comment (it gets confusing between the rust types in parquet and the thrift structs defined by the parquet thrift IDL :)). But I guess you could just have the column chunk as the bare minimum of just an offset and path and then just do decoding purely based on the data page headers. -- 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]
