[
https://issues.apache.org/jira/browse/NIFI-2632?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15433539#comment-15433539
]
ASF GitHub Bot commented on NIFI-2632:
--------------------------------------
Github user YolandaMDavis commented on the issue:
https://github.com/apache/nifi/pull/919
@mattyb149 Understood on the above comment. Just to note it (for maybe
future refactor) it may be cool to have some abstract SplitProcessor that has
attributes, relationships and some of the split/original flow file handling
that seems to be similar across many of the split processors.
This looks good tested with flows using json and xml. Splits were executed
as expected. Extracted attributes in both cases and confirmed new attributes
for flow files transferred to split queue (original files did not contain
attributes). Attributes were also available for expression languages.
+1
I'll merge this in shortly
> Add fragment attributes for SplitJson and SplitXml
> --------------------------------------------------
>
> Key: NIFI-2632
> URL: https://issues.apache.org/jira/browse/NIFI-2632
> Project: Apache NiFi
> Issue Type: Improvement
> Components: Extensions
> Reporter: Matt Burgess
> Assignee: Matt Burgess
>
> Some "splitting" processors such as SplitText and SplitContent write
> attributes to the split flow files indicating their fragment index, the total
> count, and the filename of the original flow file. This is done to support a
> form of "micro-batching" and/or a split-join pattern in a data flow (i.e.
> split the original file, do work on the individual pieces, and possibly merge
> them together later).
> For consistency and capability, the SplitJson and SplitXml processors should
> write these same attributes for their split flow files.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
