On Fri, 13 Dec 2019 at 20:45, Jonah H. Harris <jonah.har...@gmail.com>
wrote:
> On Fri, Dec 13, 2019 at 11:56 AM Ryan Skraba <r...@skraba.com> wrote:
>
>> I think the spec is OK with it. We've even used it in the Java API
>
>
> Hmm. I understood it as fields were required. Though, I could see how it’s
> written could also mean zero. We actually ran into an issue with one of the
> of the encoders because of a record with zero fields.
>
I had understood the "optional" or "required" qualifiers in the Avro spec
applied to the fields themselves, not to their content. For example, I
presume it would be valid to have an empty string as the value for a "doc"
field.
I may well be wrong though. It would be nice to be a little more precise
about this, wouldn't it? :)
Is there an Avro test suite anywhere?
>
If you *do* find an Avro test suite (ideally expressed as a
language-agnostic data set), please let me know! I'd find that very useful.
cheers,
rog.
