I've started looking into leveraging the Avro schema definition
support for Cascading workflows.
Currently I manually generate a wrapper class (what we call a "Datum")
for a Cascading Tuple, that gives me stronger typing and support for
array/map types, which aren't natively supported by Cascading.
But when a Tuple (a record) has many fields, generating this wrapper
class becomes painful and error-prone.
So I wanted to extend the SpecificCompiler class in Avro, but alter
what it emits to be my target Java source for Datums.
However the SpecificCompiler class isn't set up to be extended
(private fields/methods), so I'm wondering if there's a better
approach that I'm missing.
e l a s t i c w e b m i n i n g