stevenzwu commented on PR #7494: URL: https://github.com/apache/iceberg/pull/7494#issuecomment-1538652434
> @stevenzwu: I like that with a correct reader we can make sure that the `IcebergSource` could read whatever data structure we want. It would be good to have the possibility for the same thing for the `FlinkSink` (or IcebergSink later). Wouldn't this PR close this possibility by forcing the Sink to use `RowData`? @pvary `FlinkSink` can still support other input types (like Avro GenericRecord). The behavior (internally) is that all input type is converted to Flink `RowData`, which is the only supported record format for the underline file writer in Iceberg. That is the reason why @huyuanfeng2018 was suggesting if `RowData` is[ the only key type we need to support](https://github.com/apache/iceberg/issues/7393). If in the future, Flink sink supports other record format for the underline file writer, maybe we can revisit this decision. for now, `RowData` is the only type Flink file writer supports. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
