Aw: Re: SplitRecord vs SplitText: fragment attributes

[Uwe Geercken]

James,

 

added Issue 3988.

 

Rgds,

 

Uwe

 

Gesendet: Freitag, 26. Mai 2017 um 17:05 Uhr
Von: "James Wing" <jvw...@gmail.com>
An: users@nifi.apache.org
Betreff: Re: SplitRecord vs SplitText: fragment attributes

Uwe,
 

Please do create a JIRA.  I agree with you that SplitRecord should provide compatible fragment.* attributes like the older split processors.  Partially for a consistent user experience, and partly for compatibility with MergeContent and other processors that read the fragment.* attributes.
 

Thanks,
 

James

 

On Fri, May 26, 2017 at 7:17 AM, Uwe Geercken <uwe.geerc...@web.de> wrote:

Hello,

 

I have used both the SplitRecord and SplitText processors. When using the SplitText processor, the flowfile gets various attributes for the fragment which in turn can be used to generate a unique filename for the output with e.g. PutFile. I was using the fragment.index attribute for this.

 

The SplitRecord does not create attributes for the fragments. The only attribute I can use to generate a unique filename would be the uuid of the flowfile. While this is ok for me, I would recommend that the SplitRecord would generate similar attributes for the fragments as the SplitText processor. And btw: the other SplitXXX processors also have attributes for the fragments.

 

Should I open a jira on this?

 

Rgds,

 

Uwe