Sure, you can use a non-cached state. However, if you write code like
below for a non-cached state, and learn about caching later on, and
think, caching is a cool feature, I want to use it, you would simply
want to enable caching (without breaking your code).

Processor code should always work independently if caching is enabled or
not.

-Matthias

On 09/04/2016 06:56 PM, Eno Thereska wrote:
> Hi Matthias,
> 
> Thanks for the good questions. 
> 
> There is still the option of not using cached state. If one uses cached state 
> it will dedup for stores and forwarding further. But you can always disable 
> caching and do what you say.
> 
> Eno
> 
>> On 4 Sep 2016, at 17:36, Matthias J. Sax <matth...@confluent.io> wrote:
>>
>> Sorry for not being precise. What I meant be "completely" is for a
>> single processor. Assume I want to have the following pattern:
>>
>>  process(...) {
>>    if (someCondition) {
>>      state.put(...)
>>      context.forward(...);
>>    } else {
>>      context.forward(...);
>>  }
>>
>> Ie, for some record I do update the state and emit output records, for
>> other records I only emit output records. This work in current design.
>> However, if a "cached state" would be used, it would not work any more.
>>
>>
>> -Matthias
>>
>> On 09/04/2016 05:58 PM, Damian Guy wrote:
>>> Hi Matthias,
>>>
>>> Thanks for bringing the conversation across to the thread.
>>>
>>> I think a main limitation would be, that you cannot mix the 4 patterns
>>>> within a single application anymore (iff you use a "caches state"). If
>>>> you have processor with a "cached state" this disables direct usage of
>>>> context.forward() completely -- if I understand the design correctly.
>>>> Thus, if a "cached state" is used, forwarding is only possible via state
>>>> updates.
>>>>
>>>>
>>> The above statement is not correct. Caching doesn't completely disable
>>> forwarding, it only disables it for Processors that are using State Stores.
>>> In all other cases context.forward() works as it does now.
>>>
>>> Thanks,
>>> Damian
>>>
>>
> 

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to