[
https://issues.apache.org/jira/browse/NIFI-1234?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15035772#comment-15035772
]
ASF GitHub Bot commented on NIFI-1234:
--------------------------------------
Github user olegz commented on the pull request:
https://github.com/apache/nifi/pull/136#issuecomment-161287833
I am on the side of @apiri. I can't imagine that was the desirable behavior
either, since it was simply inconsistent. If one **explicitly** requests an
array, than it should return an array, no matter how many elements in it.
> Inconsistent container behavior in ConvertAvroToJSON
> ----------------------------------------------------
>
> Key: NIFI-1234
> URL: https://issues.apache.org/jira/browse/NIFI-1234
> Project: Apache NiFi
> Issue Type: Bug
> Components: Extensions
> Affects Versions: 0.4.0
> Reporter: Douglas Doughty
> Assignee: Aldrin Piri
> Priority: Minor
>
> ConvertAvroToJSON has 2 container options: array and none.
> When set to 'array', usage dictates an "array of objects" should be produced.
> But, if only a single Avro record is passed in, an array is not created--a
> JSON object is created.
> I believe it should be wrapped in an array. This caused errors when chained
> with SplitJSON (which expects an array).
> Use case scenario: ExecuteSQL returns an unknown number of results ->
> ConvertAvroToJSON -> SplitJSON. If ExecuteSQL returns 1 result, then the
> operation fails.
> Work around: Set up a failure relationship between SplitJSON and the next
> processor--the object will be forwarded, but an error will be logged on the
> SplitJSON processor.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
