Hi Kenn,
I'll quote here all the recent comments (from this and the (closed)
thread [1]):
> Your proposed feature is sensitive to all data that is not in
timestamp order, which is not the same as late. In Beam "late" is
defined as "assigned to a window where the watermark has passed the end
On Tue, Nov 12, 2019 at 1:36 AM Jan Lukavský wrote:
> Hi,
>
> this is follow up of multiple threads covering the topic of how to (in a
> unified way) process event streams. Event streams can be characterized
> by a common property that ordering of events matter.
1. events are ordered (hence
One concern with (1) is that it may not be cheap to do for all
runners. There also seems to be the implication that in batch elements
would be 100% in order but in streaming kind-of-in-order is OK, which
would lead to pipelines being developed/tested against stronger
guarantees than are generally
Hi,
this is follow up of multiple threads covering the topic of how to (in a
unified way) process event streams. Event streams can be characterized
by a common property that ordering of events matter. The processing
(usually) looks something like
unordered stream -> buffer (per key) ->