[ 
https://issues.apache.org/jira/browse/NIFI-12554?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17804748#comment-17804748
 ] 

David Handermann commented on NIFI-12554:
-----------------------------------------

Responding to your questions in reverse order...

Regarding the transform cache, it does not necessarily need to be declared as 
volatile.

Regarding the property names, that is important to consider. The new 
migrateProperties() method allows components to be updated with new property 
names, and programmatically migrate from old property names when required. With 
that approach, it should be possible to migrate both Processors to new property 
names that use the standard Title Case formatting, removing the need for any 
prefixing or different formatting. See other Processors like InvokeHTTP for 
examples of using migrateProperties.

Regarding the validation, including the stack trace details in the validation 
message sounds acceptable. Moving to a single Jolt Specification property that 
supports either direct specification, or file references, sounds like the best 
way to go. This should may also be an opportunity to use migrateProperties.

> Refactor JoltTransformJSON and JoltTransformRecord processors in order to 
> reduce duplicate code
> -----------------------------------------------------------------------------------------------
>
>                 Key: NIFI-12554
>                 URL: https://issues.apache.org/jira/browse/NIFI-12554
>             Project: Apache NiFi
>          Issue Type: Sub-task
>            Reporter: Daniel Stieglitz
>            Assignee: Daniel Stieglitz
>            Priority: Major
>
> There is a lot of duplicate code between the JoltTransformJSON and 
> JoltTransformRecord processors. As a result each time there is a bug 
> discovered in the duplicate code there has to be a fix applied in both places 
> (e.g. NIFI-11959 and NIFI-12165).  This ticket aims to pull up the common 
> code between JoltTransformJSON and JoltTransformRecord similar to what has 
> been done for PutElastisearchJSON and PutElastisearchRecord processors with 
> the creation of AbstractPutElasticsearch.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to