[
https://issues.apache.org/jira/browse/BEAM-644?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15522966#comment-15522966
]
Aljoscha Krettek commented on BEAM-644:
---------------------------------------
[~kenn] I was referring to "two clusters of elements from two separate input
elements" but that's somewhat besides the point because I was thinking about
how a Kafka source would be implemented as a combination of {{DoFn}} plus
{{SplittableDoFn}}. There you need to manage the watermark at the
{{SplittableDoFn}} which would be responsible for reading from topics.
I think we might be talking about different things here. As I said, the
proposed changes are very good in how they simplify the API of {{DoFn}} and
also clean up stuff around allowed time skew.
What I was thinking about is in general a problem with watermarks. I though
that the proposal here was meant to fixed that but I don't think we can. What I
was trying to get at essentially boils down to this: If we want our watermark
to be 100 % correct then we can never advance it because we never know what
timestamps future elements will have. (For the general case, where any data
with any timestamp can arrive at any point in (processing) time.). I was just
pondering that and I'm afraid it derailed the discussion a bit.
> Primitive to shift the watermark while assigning timestamps
> -----------------------------------------------------------
>
> Key: BEAM-644
> URL: https://issues.apache.org/jira/browse/BEAM-644
> Project: Beam
> Issue Type: New Feature
> Components: beam-model
> Reporter: Kenneth Knowles
> Assignee: Kenneth Knowles
>
> There is a general need, especially important in the presence of
> SplittableDoFn, to be able to assign new timestamps to elements without
> making them late or droppable.
> - DoFn.withAllowedTimestampSkew is inadequate, because it simply allows one
> to produce late data, but does not allow one to shift the watermark so the
> new data is on-time.
> - For a SplittableDoFn, one may receive an element such as the name of a log
> file that contains elements for the day preceding the log file. The timestamp
> on the filename must currently be the beginning of the log. If such elements
> are constantly flowing, it may be OK, but since we don't know that element is
> coming, in that absence of data, the watermark may advance. We need a way to
> keep it far enough back even in the absence of data holding it back.
> One idea is a new primitive ShiftWatermark / AdjustTimestamps with the
> following pieces:
> - A constant duration (positive or negative) D by which to shift the
> watermark.
> - A function from TimestampedElement<T> to new timestamp that is >= t + D
> So, for example, AdjustTimestamps(<-60 minutes>, f) would allow f to make
> timestamps up to 60 minutes earlier.
> With this primitive added, outputWithTimestamp and withAllowedTimestampSkew
> could be removed, simplifying DoFn.
> Alternatively, all of this functionality could be bolted on to DoFn.
> This ticket is not a proposal, but a record of the issue and ideas that were
> mentioned.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
