Provides compatibility between Java and Scala collections, since Avro writers are written in Java and use a custom implementation of Collection, but this library also uses those writers to convert records into Array[Byte].
Provides compatibility between Java and Scala collections, since Avro writers are written in Java and use a custom implementation of Collection, but this library also uses those writers to convert records into Array[Byte].
This is a hack.
This is a hack. When a Record datum is created from a dataframe its schema namespace is defined by the location of the attribute in the dataframe schema as opposed to the avro schema that is to be imposed. This means when the logic tries to resolve the namespace of the datum (record) from the given types in the union it cannot find it. The below logic works by if all else fails trying to match the datum schema name with any of the union type names and if found returning that.
schema
data
This class redefines the way fields are retrieved from a record since they could be either an IndexedRecord (if written by Avro library) or a GenericRow, if written this library.