mkaravel commented on PR #2971: URL: https://github.com/apache/parquet-java/pull/2971#issuecomment-2778784931
> I took a pass through here to check the corner case behaviour against the C++ PR after rewriting the example files to deal more intentionally with the issues we've run into ( [apache/parquet-testing#70](https://github.com/apache/parquet-testing/pull/70) ). I think the only inconsistency left is that a completely empty column chunk here will export `xmin`/`ymin`/`xmax`/`ymax` as either all zeroes or all `NaN`s to Thrift (whatever happens if you don't set a required Thrift field). In C++ I export these as `Inf`/`Inf`/`-Inf`/`-Inf`...I still think this is the least ambiguous way to do this but any other convention you all agree (e.g., all `NaN`s or unsetting the entire BoundingBox) is fine with me. I still think that returning `(inf, -inf, inf, -inf)` will be confusing. Technically it is not a valid bounding **geometric** box because the order of the bounds is reversed. Personally I prefer returning NaNs instead (see also my comment above). -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
