Bug in NiFi XML processor

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Bug in NiFi XML processor

Sadasivan, Sreejith

Hi,

 

I have been using NiFi for a while without any issues. Recently I found a bug in the EvaluateXPath processor (I was using the 1.2.0 version of it, recently using 1.4.0). The problem is described below:

 

I have a CDATA section in my xml. All the CDATA are processed properly but the processor fails if it has [JM] at the end of the CDATA section. Please find the screen shot attached.

 

<![CDATA[This is sample text. [JM]]]> this will fail.

<![CDATA[This is sample text. [JM] ]]> providing a space after [JM] will work.

 

So I think this is a bug in the NiFi XML processors, which considers the closing bracket of [JM] as the end of CDATA.

 

Thanks,

Sreejith

Reply | Threaded
Open this post in threaded view
|

Re: Bug in NiFi XML processor

Koji Kawamura-2
Hi Sreejith,

Do you still have the issue? Unfortunately the attached screenshot is
dropped so I couldn't see what error you got.
I tried to reproduce the issue, but EvaluateXPath runs fine with your
example data regardless having whitespace or not.
Here is a flow template that I used to confirm:
https://gist.github.com/ijokarumawak/6951783fea0e02db08fc97c2d551de2b

If the issue still persists, please feel free to submit a NiFi JIRA issue.
https://issues.apache.org/jira/projects/NIFI

Thanks,
Koji

On Mon, Nov 27, 2017 at 6:16 PM, Sadasivan, Sreejith
<[hidden email]> wrote:

> Hi,
>
>
>
> I have been using NiFi for a while without any issues. Recently I found a
> bug in the EvaluateXPath processor (I was using the 1.2.0 version of it,
> recently using 1.4.0). The problem is described below:
>
>
>
> I have a CDATA section in my xml. All the CDATA are processed properly but
> the processor fails if it has [JM] at the end of the CDATA section. Please
> find the screen shot attached.
>
>
>
> <![CDATA[This is sample text. [JM]]]> this will fail.
>
> <![CDATA[This is sample text. [JM] ]]> providing a space after [JM] will
> work.
>
>
>
> So I think this is a bug in the NiFi XML processors, which considers the
> closing bracket of [JM] as the end of CDATA.
>
>
>
> Thanks,
>
> Sreejith