alamb commented on issue #2992: URL: https://github.com/apache/arrow-datafusion/issues/2992#issuecomment-1200504172
In general I think this proposal sounds like a nice compromise between "DataFusion can handle reading remote object stores out of the box" and "my project needs fine grained control" and as @tustvold mentions I think it fits well with the philosophy of DataFusion Thank you for proposing this > So for people manually constructing ParquetExec as part of a custom TableProvider, they can obviously just pass the new parameter. Yes this makes sense. > The option could then be added to ParquetFormat to allow constructing a ListingOptions and in turn a ListingTable that uses this setting. Yes this makes sense > Finally we could expose it in ParquetReadOptions so that it can be used with SessionContext::read_parquet, but I'm not sure if we want to expose it at this high a level. I suggest we start leaving it out of `ParquetReadOptions` because I think that interface is already somewhat confusing (see https://github.com/apache/arrow-datafusion/pull/2988) and if someone need the API we/they can add it -- 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]
