t I need (such as getHeadKey(),
>>> getLastl(), or makeKey(Key,Object)) to “navigate” in the key space, and a
>>> cached hash code to make hash maps faster. My generic algorithms do not
>>> need to know how many fields are exposed in the Key, but they need to be
>>
ilto:user@flink.apache.org>
Objet : Re: How to make a generic key for groupBy
Hey Arnaud!
I have made a quick sample implementation of how you can very efficiently
support generic keys, like yours. I put the code in this repository:
https://github.com/StephanEwen/flink-generic-keys
It implemen
. My generic algorithms do not
>> need to know how many fields are exposed in the Key, but they need to be
>> able to construct another key from two keys.
>>
>>
>>
>> Arnaud
>>
>>
>>
>> *De :* ewenstep...@gmail.com [mailto:ewenstep...@gmail.c
han Ewen
> *Envoyé :* vendredi 24 avril 2015 11:14
> *À :* user@flink.apache.org
> *Objet :* Re: How to make a generic key for groupBy
>
>
>
> Hi Arnaud!
>
>
>
> Thank you for the warm words! Let's find a good way to get this to work...
>
>
>
> As a bit
[mailto:ewenstep...@gmail.com] De la part de Stephan
Ewen
Envoyé : vendredi 24 avril 2015 11:14
À : user@flink.apache.org
Objet : Re: How to make a generic key for groupBy
Hi Arnaud!
Thank you for the warm words! Let's find a good way to get this to work...
As a bit of background:
In Flink, the API nee
Hi Arnaud!
Thank you for the warm words! Let's find a good way to get this to work...
As a bit of background:
In Flink, the API needs to now a bit about the types that go through the
functions, because Flink pre-generates and configures serializers, and
validates that things fit together.
It is
Will you elaborate on your use case? It would help to find out where Flink
shines. IMO, its a great project, but needs more differentiation from Spark.
On Thu, Apr 23, 2015 at 7:25 AM, LINZ, Arnaud
wrote:
> Hello,
>
>
>
> After a quite successful benchmark yesterday (Flink being about twice
> f
Hello,
After a quite successful benchmark yesterday (Flink being about twice faster
than Spark on my use cases), I’ve turned instantly from spark-fan to flink-fan
– great job, committers!
So I’ve decided to port my existing Spark tools to Flink. Happily, most of the
difficulty was renaming clas
