[
https://issues.apache.org/jira/browse/NIFI-4181?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16130856#comment-16130856
]
ASF GitHub Bot commented on NIFI-4181:
--------------------------------------
Github user markap14 commented on the issue:
https://github.com/apache/nifi/pull/2003
@Wesley-Lawrence sorry - i must have missed the email that you'd commented
on the PR. Sorry I didn't notice until now. So I think that what you're
proposing here is that we should have a property named "Schema Text Format" on
the readers/writers, in addition to "Schema Text", and the Schema Text Format
would tell the service how to parse the "Schema Text." For example, there would
be two options: "Avro" and "String Column Names". In addition to this, we could
potentially also introduce a new Schema Registry as I described above. This
means that for one-off types of schemas, such as when we have a CSV file and we
only need to use the schema once, we don't have to go to the trouble of
creating a Schema Registry and adding entries to it - we'd just set the column
names (optionally using Expression Language) in the Reader. And we would add
this capability to JSON, etc. as well.
Does that all sound accurate? Just want to make sure that we are on the
same page here.
> CSVReader and CSVRecordSetWriter services should be able to work given an
> explicit list of columns.
> ---------------------------------------------------------------------------------------------------
>
> Key: NIFI-4181
> URL: https://issues.apache.org/jira/browse/NIFI-4181
> Project: Apache NiFi
> Issue Type: Improvement
> Reporter: Wesley L Lawrence
> Priority: Minor
> Attachments: NIFI-4181.patch
>
>
> Currently, to read or write a CSV file with *Record processors, the CSVReader
> and CSVRecordSetWriters need to be given an avro schema. For CSV, a simple
> column definition can also work.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
