When using kafka-avro-console-producer, the only way that we allow users to describe the schema of their data is by a literal representation, with --property value.schema="...".
Users will often already have their schema in a file. Rather than forcing them to inline it, the script should take a --property value.schema.file=... argument that reads in the schema directly.
What about also value.schema.url and point at an existing schema in the registry?
That also seems useful. I think you'd need both since sometimes you're producing data for the first time -- before it's ever registered its schema with SR.
value.schema.file now exists as an option
value.schema.filenow exists as an option
What version of Confluent is this available in?
Also, is/will there be a separate value.schema.file for the key and value separately?
Also... where is the doco on this? :) I can't find it in Confluent's docination.
Thanks!
Most helpful comment
value.schema.filenow exists as an option