HeartSaVioR commented on issue #23854: [SPARK-22000][SQL] Inject necessary casts when data type of columns and matching type of fields are not equal URL: https://github.com/apache/spark/pull/23854#issuecomment-466447468 It turns out to be more complicated than expected. I need to apply projection with cast only when deserialization happens, but some of cases I had to play with physical plan which I had to rebuild deserializer. After the fix, nothing leverages `Dataset.deserializer` so I've removed it. According to comment, in sample case Spark should do type check and provide error, but as we've seen the error is not from type check so I'm wondering eagerly initializing `deserializer` does the type check properly.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
