Hi Till,
sorry for this. I sent this e-mail from my other account, which is not
subscribed to this thread and according to the website should have been
rejected. Once I realized it I sent another e-mail from the correct account.
Bestt,
Daniel
Till Rohrmann ezt írta (időpont: 2018. okt. 8., H,
9
Hi Daniel,
Didn't you post this question before? Let's not spread the discussion out
over multiple threads.
Cheers,
Till
On Sun, Oct 7, 2018 at 4:27 PM Dániel Berecz wrote:
> Hi everyone,
>
> I would like to ask if there is any conceptual problem with having MapState
> available for operator s
Hi everyone,
I would like to ask if there is any conceptual problem with having MapState
available for operator states, or is it just a lack of implementation?
Best regards,
Daniel Berecz
One challenge would be duplicate keys in this context.
Am Do., 4. Okt. 2018 um 10:17 Uhr schrieb Till Rohrmann <
trohrm...@apache.org>:
> Hi Daniel,
>
> I don't think that there is a fundamental problem of having MapState
> available for operator state. First, there are some questions to be
> ans
Hi Daniel,
I don't think that there is a fundamental problem of having MapState
available for operator state. First, there are some questions to be
answered though: How do you union map state and how do you split map state
in case of repartitioning. Once this has been answered one needs to
impleme
Hi everyone,
I would like to ask if there is any conceptual problem with having MapState
available for operator states, or is it just a lack of implementation?
Best regards,
Daniel Berecz
