Speaking of Elasticsearch, are you using InvokeHttp for the bulk insert, or
something else? There is a Jira case (NIFI-1275) to add Elasticsearch
processor(s) to NiFi, we welcome all thoughts and comments there on how to make
things easier.
Regards,
Matt
Sent from my iPhone
> On Dec 13, 2015, at 6:57 PM, Aldrin Piri <aldrinp...@gmail.com> wrote:
>
> You are correct in that it takes the literal representation of what is
> entered. There are two ways to handle this:
>
> 1. With delimiter strategy of 'Text': Hit shift + enter to provide a new
> line
> 2. With delimiter strategy of 'Filename': Provide a path to a file that
> is only a newline character
>
> With regards to your question on leaving the footer empty, it should work
> as you anticipate for the Merge Format of 'Binary Concatenation.' Making a
> quick test flow, I have verified this, but please let us know if it is not
> working for your particular configuration.
>
> On Sun, Dec 13, 2015 at 6:18 PM, Igor Kravzov <ikrav...@brandprotect.com>
> wrote:
>
>> I want to use MergeContent processor to merge tweets to bulk insert into
>> Elasticsearch index. For this I need command and tweets (each) to be
>> separated by \n.
>>
>> This is how it should look like
>> { action: { metadata }}\n
>> { request body }\n
>>
>> Which will be
>> { "index" )\n
>> { tweet1 }\n
>> { twee2 }\n
>>
>> When I put \n as demarcator the processor actually adds \n as a string
>> instead of new line separator. Is it possible to make it actual new line?
>> Also is it possible to leave or make footer empty?
>>
>> Thanks in advance.
>>
