scovich commented on PR #8324: URL: https://github.com/apache/arrow-rs/pull/8324#issuecomment-3282060226
> > > > CON: Architectural violation to make parquet-variant crate (at least somewhat) aware of parquet-variant-compute crate that depends on it. > > > > > > > > > Another potential option is to move some/all the specialized builder code into the parquet-variant crate 🤔 > > > > > > Moving `VariantArrayBuilder` to from parquet-variant-compute to parquet-variant crate would indeed resolve the architectural violation. Is that even such a bad thing? > > It doesn't seem like a bad idea to me, to be honest. > > The only potential issue would be people who wanted to use parquet-variant without the dependency on arrow (I am only theorizing here, I don't know if that is actually an important usecase). That's an interesting point. In https://github.com/delta-io/delta-kernel-rs/, for example, arrow is the default (but not required) engine implementation -- a Delta connector based on the kernel could choose to use its own native facilities instead. But variant is part of the Delta specification, so it could be important to have access to a robust binary variant implementation even if the engine isn't otherwise using arrow. -- 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: github-unsubscr...@arrow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
