lidavidm opened a new issue, #924:
URL: https://github.com/apache/arrow-adbc/issues/924

   All the drivers in a language need to follow the same specifications, the 
same conventions, and have a lot of similar scaffolding. For maintainability's 
and testing's sake it may make more sense to create base drivers that handle 
most of the ADBC state and API boundaries and just dispatch to a 
vendor-specific backend to do things like query execution. We only have SQLite 
and PostgreSQL for now, but I expect we'll have an ODBC adapter soon. And it 
seems like the drivers may want more similar functionality than we expect: 
PostgreSQL decimals, for instance, will likely need type inference in the same 
way all SQLite type do, if we want to actually convert them to decimals.
   
   It would also make it easier to integrate observability features like 
detailed logging, OpenTelemetry support, and so on, if we can (mostly) handle 
them in one place.
   
   Admittedly, a lot of the code is actually in the query execution part, which 
may not be (easily) shareable. So it may not be as easy of a win as I think.


-- 
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]

Reply via email to