Github user JohannesDaniel commented on the issue:
https://github.com/apache/nifi/pull/2587
@markap14 @pvillard31
- I refactored some code as the cases (coerce==true && drop==false) and
(coerce==false && drop==true) in some cases showed an unexpected behavior
- Data like <tag>content<subtag>content</subtag>content</tag> now can be
parsed
- Maps (e. g.
<map_field><key1>value1</key1><key2>value2</key2></map_field>) are now supported
- The reader is now able to parse single records (e. g.
<record><field1/><field2/></record>) as well as arrays of records (e. g.
<root><record/><record/></root>). I added a property to make it configurable
whether the reader shall expect a single record or an array. One question: As
there are only two options for this, I defined AllowableValues for this
property. Despite that, I think it would be reasonable to enable EL for this
property. But how can this be realized?
- I removed the root validation, but remained the check for record tag
names in order to support processing data like this
<root><record/><record/><other/></root> (tag "other" will be ignored if check
for record tag name is activated)
---
