Gang Wu
Gang Wu
Is it possible to use `std::basic_string_view` ~~or `std::basic_string_view`~~?
It doesn't make a big difference to change private members to protected or public as we enforce backward compatibility for all public APIs. BTW, is it possible to introduce similar...
If you do need to cache (compressed or uncompressed) pages, I think it is unavoidable to implement some sort of page cache. Do you want to contribute this?
It would be good to make it configurable to set different cache strategies including compressed/uncompressed, requested columnIds, eviction rule, etc. For the cache key, my idea is to suggest a...
I agree with @gszadovszky. Let's start simple.
> I'm just noticing this PR. Is there a reason to make this behavior optional instead of standard? Before proceeding, I think we need to agree on this.
Sure, I will take a look. Thanks!
Thanks for the review @mkaravel @paleolimbot @jiayuasu! Let me clarify some things about Parquet statistics first: - Parquet does not consider `NaN` when collecting min/max values for float/double values. So...
> I wasn't able to find a null count for a row group in statistics for all null values (or otherwise) because (at least in C++) the statistics aren't written...
@mkaravel Thanks for the input! Let's discuss this topic in the format PR.