Angel,

OK, sounds like we've got a good path forward. I created a JIRA for this [1].

Is this by chance something that you would be able / interested in working on?

Thanks
-Mark

[1] https://issues.apache.org/jira/browse/NIFI-2161 
<https://issues.apache.org/jira/browse/NIFI-2161>



> On Jul 1, 2016, at 9:10 AM, Angel Kafazov <[email protected]> wrote:
> 
> Hi Mark,
> 
> Yes, I think this option would be exactly what I am looking for. It would 
> make sense to add it, as there doesn't seem to be any other way around this 
> using any of the available processors.
> 
> Best Regards,
> Angel
> 
> 
> On Fri, Jul 1, 2016 at 4:01 PM, Mark Payne <[email protected] 
> <mailto:[email protected]>> wrote:
> Ah, I did miss that. In the cases that I have used it, an empty string would 
> not be a valid value in the JSON, so I
> have been able to route on whether or not the value is 'null or empty' by 
> checking ${myAttribute:isEmpty()}.
> Would it help if there were a new option in this Processor for the "Null 
> Value Representation" such that if the value
> does not exist, no attribute is even added? Then you'd be able to simply 
> check for the existence of the attribute.
> 
> Thanks
> -Mark
> 
>> On Jul 1, 2016, at 8:08 AM, Angel Kafazov <[email protected] 
>> <mailto:[email protected]>> wrote:
>> 
>> Hi Mark,
>> 
>> The EvaluateJsonPath always sets the attribute with empty string value, 
>> regardless if the JSON field exists or not. It is only possible to check if 
>> the field has a empty value or not, but not possible to check if it actually 
>> exists. Am I missing something?
>> 
>> Best Regards,
>> Angel
>> 
>> 
>> On Thu, Jun 30, 2016 at 7:31 PM, Mark Payne <[email protected] 
>> <mailto:[email protected]>> wrote:
>> Hi Angel,
>> 
>> I don't know the reasoning behind only using 'unmatched' for FlowFile 
>> Content, but I am guessing
>> it has to do with the idea that the FlowFile is not modified when routed to 
>> this relationship and the
>> author probably wanted to avoid the ambiguity that would occur when some 
>> paths matched but
>> others didn't.
>> 
>> That being said, the pattern that is typically followed is to use 
>> EvaluateJsonPath with a destination
>> of attribute and then to follow this processor with a RouteOnAttribute 
>> processor to check if all of the
>> attributes were added.
>> 
>> Thanks
>> -Mark
>> 
>> 
>>> On Jun 30, 2016, at 12:05 PM, Angel Kafazov <[email protected] 
>>> <mailto:[email protected]>> wrote:
>>> 
>>> Hello,
>>> 
>>> I am considering using EvaluateJsonPath processor to check if JSON 
>>> formatted FlowFile content contains a number of predefined fields. The 
>>> processor defines three relationships: failure, unmatched and matched as 
>>> follows:
>>> 
>>> failure     FlowFiles are routed to this relationship when the JsonPath 
>>> cannot be evaluated against the content of the FlowFile; for instance, if 
>>> the FlowFile is not valid JSON
>>> unmatched   FlowFiles are routed to this relationship when the JsonPath 
>>> does not match the content of the FlowFile and the Destination is set to 
>>> flowfile-content
>>> matched     FlowFiles are routed to this relationship when the JsonPath is 
>>> successfully evaluated and the FlowFile is modified as a result
>>> 
>>> If the flowfile-content Destination is used, the existance only one field 
>>> can be checked. To validate multiple fields, flowfile-attribute destination 
>>> must be used, this means however that the unmatched relationship will not 
>>> be used.
>>> 
>>> This actually prevents one from being able to check existance multiple JSON 
>>> fields in the content, if I understand correctly. What is the reason for 
>>> this implementation? Is there any other processor that could be used to 
>>> achieve the task?
>>> 
>>> Regards,
>>> Angel
>>> 
>>> 
>>> Angel Kafazov
>>> CST LTD
>>> Bul. Bulgaria, blok 10, 6500 Svilengrad
>>> Bulgaria
>>> +359877336002 <tel:%2B359877336002> (BG Mobile)
>>> +4961509729845 <tel:%2B4961509729845> (DE Landline)
>>> Email: [email protected] <mailto:[email protected]>
>>> http://cst-bg.net <http://cst-bg.net/>
>> 
> 
> 

Reply via email to