Hey Frederik,
I took a look at the template. The ConvertRecord processor is configured with
the CSVRecordSetWriter.
The problem appears to be that the writer is configured with the "Schema Access
Strategy" of "Use 'Schema Text' Property"
and the Schema Text property is set to "${avro.schema}". As a result, it is
expecting a FlowFile attribute to be present
named "avro.schema" but that attribute is not present. Unfortunately, in
version 1.3.0 of NiFi you will need to use a Schema Registry
and provide the schema for the data. We knew that this was a bit of a pain
point, so in the master branch, we have provided
a new option on the writers, so that you can set the Schema Access Strategy to
"Inherit Record Schema." This would allow you
to use the schema that the Record Reader determines. This works far more
seamlessly in this case because the schema is already
present in the data.
Unfortunately, though, I think that for now you would either have to build
'master' and use that, or you would have to provide the schema
via UpdateAttribute (to add the avro.schema attribute) or use a Schema Registry.
Hope this helps!
-Mark
> On Aug 7, 2017, at 1:17 AM, Frederik <[email protected]> wrote:
>
> Hey,
>
> no problem. Please find the template file below. Let me know if I can
> provide anything else. thanks
>
> convert_record_template.xml
> <http://apache-nifi-developer-list.39713.n7.nabble.com/file/n16590/convert_record_template.xml>
>
>
>
>
> --
> View this message in context:
> http://apache-nifi-developer-list.39713.n7.nabble.com/Avro-Reader-reading-from-FlowFile-throws-SchemaNotFoundException-tp16574p16590.html
> Sent from the Apache NiFi Developer List mailing list archive at Nabble.com.
