Venkat,

You are correct that right now the incoming message delimiter is hard-coded
as "/n".

The property you see in the UI for message delimiter is actually for the
output of the processor. It is used when you increase the batch size > 1
and it writes multiple messages to a single flow file, it uses the value of
that property to separate them.

We definitely want to expose the incoming delimiter as something that can
be set through a property on the processor. I think one reason it hasn't
been done yet is because there are several strategies we'd like to support:

- exact match - this would be your case where you specify "$"
- pattern match - this would be reading until a pattern is seen to help
capture multi-line log messages that start with date patterns
- size match - this would be reading until a specified number of bytes have
been read

That being said, we shouldn't need to do all of these at once so I created
this JIRA for your scenario:

https://issues.apache.org/jira/browse/NIFI-1985

-Bryan



On Wed, Jun 8, 2016 at 2:25 AM, Venkatesh Nandigam <
[email protected]> wrote:

> Hi Team,
>
> We started using NIFI data flow in our current project by replacing node js
> tcp listeners. we are using nifi listenTCP processor.
>
> our use case: we have some devices they will send message to tcp port, from
> nifi we have to get message and then place data into kafka.
>
> Processors we are using:
>
> ListenTCP processor: to receive data from port and send to kafka topic
>
> PutKafka: place data into kafka topic.
>
> Problem we are facing:
>
> from device side we have message delimiter as "$" but nifi listenTCP
> processor is accepting only "/n". we changed the delimeter in nifi admin
> its not reflecting.. by looking at nifi code we saw TCP_DELIMETER field as
> final. then we changed that class with our delimiter its working fine..
>
> Path:
>
> https://github.com/apache/nifi/blob/master/nifi-commons/nifi-processor-utilities/src/main/java/org/apache/nifi/processor/util/listen/handler/socket/SocketChannelHandler.java
>
> Is this expected behavior? or we missing something?.. if it is excepted
> behavior is they any chance to change that modifiable field?.
>
> Thanks,
> Venkat
>

Reply via email to