Matt Burgess created NIFI-5213:

             Summary: Allow AvroReader with explicit schema to read files with 
embedded schema
                 Key: NIFI-5213
             Project: Apache NiFi
          Issue Type: Improvement
          Components: Extensions
            Reporter: Matt Burgess

AvroReader allows the choice of schema access strategy from such options as Use 
Embedded Schema, Use Schema Name, Use Schema Text, etc. If the incoming Avro 
files will have embedded schemas, then Use Embedded Schema is best practice for 
the Avro Reader. However it is not intuitive that if the same schema that is 
embedded in the file is specified by name (using a schema registry) or 
explicitly via Schema Text, that errors can occur. This has been noticed in 
QueryRecord for example, and the error is also not intuitive or descriptive (it 
is often an ArrayIndexOutOfBoundsException).

To provide a better user experience, it would be an improvement for AvroReader 
to be able to successfully process Avro files with embedded schemas, even when 
the Schema Access Strategy is not "Use Embedded Schema". Of course, the 
explicit schema would have to match the embedded schema, or an error would be 
reported (and rightfully so).

This message was sent by Atlassian JIRA

Reply via email to