Jordan Sammut created NIFI-15209:
------------------------------------
Summary: JoltTransformRecord should not only take schema from
first record
Key: NIFI-15209
URL: https://issues.apache.org/jira/browse/NIFI-15209
Project: Apache NiFi
Issue Type: Bug
Reporter: Jordan Sammut
There looks to be an issue with the JoltTransformRecord processor, whereby the
schema is only being determined using the first record in a batch. This in turn
might result in inconsistencies with the Jolt and output incorrect results.
The following is a simple example:
[ \{ "operation": "shift", "spec": { "test_field": { "": "TRASH", "*": { "$":
"value.test_field" } } } } ]
Taking the above jolt, and the following input:
{ "test_field" : "" } \{ "test_field" : "value2" }
is resulting in the following result which should not be the case:
{"remove":null} \{"remove":null}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
