Hi Micah,
Le 24/06/2026 à 19:50, Micah Kornfield a écrit :
However, I think the
version number should be treated as informational only (i.e. it shouldn't
be validated by readers) and having a fine grained enumeration of features
a reader needs to understand is a better approach for the following reasons:
1. It allows for readers to produce finer grained error messages for the
options it doesn't support (for the 1 time cost of changing the magic
number, which will produce "not a parquet file")
I don't understand what problem this is solving that is not already
solved, as readers are already able to emit a fine-grained error message
when they encounter an unsupported feature.
As an additional data point, we already added something like this in
Arrow and I don't know of an implementation that uses it:
https://github.com/apache/arrow/blob/cffe22d6fe7f73836d2d9177b4b4fb2d1fcf4d1f/format/Schema.fbs#L54-L81
(it was added in https://github.com/apache/arrow/pull/7502 and you
probably remember who the author was :-D)
Regards
Antoine.
