Is your feature request related to a problem or challenge? Please describe what you are trying to do.

Currently, the reader schema cannot be created with knowledge of the actual writer schema of the Avro file. Access to the discovered schema would more flexibly support complex resolution and projection use cases that are not covered by the Avro specification. One important instance is compatibility with Spark, which performs custom name-based resolution based on the Avro writer schema and the output schema of the operator, lacking record type names to resolve against.

Describe the solution you'd like

Since the Avro file header is already read as part of the construction of the reader to resolve the reader schema and apply the projection, it's a small change to expose this step in the builder API.