You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@flink.apache.org by Aljoscha Krettek <al...@apache.org> on 2016/05/02 10:29:12 UTC
Re: Multiple windows with large number of partitions
Hi,
what do you mean by "still experiencing the same issues"? Is the key count
still very hight, i.e. 500k windows?
For the watermark generation, specifying a lag of 2 days is very
conservative. If the watermark is this conservative I guess there will
never arrive elements that are behind the watermark, thus you wouldn't need
the late-element handling in your triggers. The late-element handling in
Triggers is only required to compensate for the fact that the watermark can
be a heuristic and not always correct.
Cheers,
Aljoscha
On Thu, 28 Apr 2016 at 21:24 Christopher Santiago <ch...@ninjametrics.com>
wrote:
> Hi Aljoscha,
>
>
> Aljoscha Krettek wrote
> >>is there are reason for keying on both the "date only" field and the
> "userid". I think you should be fine by just specifying that you want 1-day
> windows on your timestamps.
>
> My mistake, this was from earlier tests that I had performed. I removed it
> and went to keyBy(2) and I am still experiencing the same issues.
>
>
> Aljoscha Krettek wrote
> >>Also, do you have a timestamp extractor in place that takes the timestamp
> from your data and sets it as the internal timestamp field.
>
> Yes there is, it is from the BoundedOutOfOrdernessGenerator example:
>
> public static class BoundedOutOfOrdernessGenerator implements
> AssignerWithPeriodicWatermarks<Tuple3<DateTime, String, String>> {
> private static final long serialVersionUID = 1L;
> private final long maxOutOfOrderness =
> Time.days(2).toMilliseconds();
> private long currentMaxTimestamp;
>
> @Override
> public long extractTimestamp(Tuple3<DateTime, String, String>
> element, long previousElementTimestamp) {
> long timestamp = element.f0.getMillis();
> currentMaxTimestamp = Math.max(timestamp, currentMaxTimestamp);
> return timestamp;
> }
>
> @Override
> public Watermark getCurrentWatermark() {
> return new Watermark(currentMaxTimestamp - maxOutOfOrderness);
> }
> }
>
> Thanks,
> Chris
>
>
>
> --
> View this message in context:
> http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Multiple-windows-with-large-number-of-partitions-tp6521p6562.html
> Sent from the Apache Flink User Mailing List archive. mailing list archive
> at Nabble.com.
>
Re: Multiple windows with large number of partitions
Posted by Aljoscha Krettek <al...@apache.org>.
Yes, please go ahead. That would be helpful.
On Mon, 2 May 2016 at 21:56 Christopher Santiago <ch...@ninjametrics.com>
wrote:
> Hi Aljoscha,
>
> Yes, there is still a high partition/window count since I have to keyby
> the userid so that I get unique users. I believe what I see happening is
> that the second window with the timeWindowAll is not getting all the
> results or the results from the previous window are changing when the
> second window is running. I can see the date/unique user count increase
> and decrease as it is running for a particular day.
>
> I can share the eclipse project and the sample data file I am working off
> of with you if that would be helpful.
>
> Thanks,
> Chris
>
> On Mon, May 2, 2016 at 12:55 AM, Aljoscha Krettek [via Apache Flink User
> Mailing List archive.] <[hidden email]
> <http:///user/SendEmail.jtp?type=node&node=6626&i=0>> wrote:
>
>> Hi,
>> what do you mean by "still experiencing the same issues"? Is the key
>> count still very hight, i.e. 500k windows?
>>
>> For the watermark generation, specifying a lag of 2 days is very
>> conservative. If the watermark is this conservative I guess there will
>> never arrive elements that are behind the watermark, thus you wouldn't need
>> the late-element handling in your triggers. The late-element handling in
>> Triggers is only required to compensate for the fact that the watermark can
>> be a heuristic and not always correct.
>>
>> Cheers,
>> Aljoscha
>>
>> On Thu, 28 Apr 2016 at 21:24 Christopher Santiago <[hidden email]
>> <http:///user/SendEmail.jtp?type=node&node=6601&i=0>> wrote:
>>
>>> Hi Aljoscha,
>>>
>>>
>>> Aljoscha Krettek wrote
>>> >>is there are reason for keying on both the "date only" field and the
>>> "userid". I think you should be fine by just specifying that you want
>>> 1-day
>>> windows on your timestamps.
>>>
>>> My mistake, this was from earlier tests that I had performed. I removed
>>> it
>>> and went to keyBy(2) and I am still experiencing the same issues.
>>>
>>>
>>> Aljoscha Krettek wrote
>>> >>Also, do you have a timestamp extractor in place that takes the
>>> timestamp
>>> from your data and sets it as the internal timestamp field.
>>>
>>> Yes there is, it is from the BoundedOutOfOrdernessGenerator example:
>>>
>>> public static class BoundedOutOfOrdernessGenerator implements
>>> AssignerWithPeriodicWatermarks<Tuple3<DateTime, String, String>> {
>>> private static final long serialVersionUID = 1L;
>>> private final long maxOutOfOrderness =
>>> Time.days(2).toMilliseconds();
>>> private long currentMaxTimestamp;
>>>
>>> @Override
>>> public long extractTimestamp(Tuple3<DateTime, String, String>
>>> element, long previousElementTimestamp) {
>>> long timestamp = element.f0.getMillis();
>>> currentMaxTimestamp = Math.max(timestamp,
>>> currentMaxTimestamp);
>>> return timestamp;
>>> }
>>>
>>> @Override
>>> public Watermark getCurrentWatermark() {
>>> return new Watermark(currentMaxTimestamp -
>>> maxOutOfOrderness);
>>> }
>>> }
>>>
>>> Thanks,
>>> Chris
>>>
>>>
>>>
>>> --
>>> View this message in context:
>>> http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Multiple-windows-with-large-number-of-partitions-tp6521p6562.html
>>> Sent from the Apache Flink User Mailing List archive. mailing list
>>> archive at Nabble.com.
>>>
>>
>>
>> ------------------------------
>> If you reply to this email, your message will be added to the discussion
>> below:
>>
>> http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Multiple-windows-with-large-number-of-partitions-tp6521p6601.html
>> To unsubscribe from Multiple windows with large number of partitions, click
>> here.
>> NAML
>> <http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>>
>
>
> ------------------------------
> View this message in context: Re: Multiple windows with large number of
> partitions
> <http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Multiple-windows-with-large-number-of-partitions-tp6521p6626.html>
> Sent from the Apache Flink User Mailing List archive. mailing list archive
> <http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/> at
> Nabble.com.
>
Re: Multiple windows with large number of partitions
Posted by Christopher Santiago <ch...@ninjametrics.com>.
Hi Aljoscha,
Yes, there is still a high partition/window count since I have to keyby the
userid so that I get unique users. I believe what I see happening is that
the second window with the timeWindowAll is not getting all the results or
the results from the previous window are changing when the second window is
running. I can see the date/unique user count increase and decrease as it
is running for a particular day.
I can share the eclipse project and the sample data file I am working off
of with you if that would be helpful.
Thanks,
Chris
On Mon, May 2, 2016 at 12:55 AM, Aljoscha Krettek [via Apache Flink User
Mailing List archive.] <ml...@n4.nabble.com> wrote:
> Hi,
> what do you mean by "still experiencing the same issues"? Is the key count
> still very hight, i.e. 500k windows?
>
> For the watermark generation, specifying a lag of 2 days is very
> conservative. If the watermark is this conservative I guess there will
> never arrive elements that are behind the watermark, thus you wouldn't need
> the late-element handling in your triggers. The late-element handling in
> Triggers is only required to compensate for the fact that the watermark can
> be a heuristic and not always correct.
>
> Cheers,
> Aljoscha
>
> On Thu, 28 Apr 2016 at 21:24 Christopher Santiago <[hidden email]
> <http:///user/SendEmail.jtp?type=node&node=6601&i=0>> wrote:
>
>> Hi Aljoscha,
>>
>>
>> Aljoscha Krettek wrote
>> >>is there are reason for keying on both the "date only" field and the
>> "userid". I think you should be fine by just specifying that you want
>> 1-day
>> windows on your timestamps.
>>
>> My mistake, this was from earlier tests that I had performed. I removed
>> it
>> and went to keyBy(2) and I am still experiencing the same issues.
>>
>>
>> Aljoscha Krettek wrote
>> >>Also, do you have a timestamp extractor in place that takes the
>> timestamp
>> from your data and sets it as the internal timestamp field.
>>
>> Yes there is, it is from the BoundedOutOfOrdernessGenerator example:
>>
>> public static class BoundedOutOfOrdernessGenerator implements
>> AssignerWithPeriodicWatermarks<Tuple3<DateTime, String, String>> {
>> private static final long serialVersionUID = 1L;
>> private final long maxOutOfOrderness =
>> Time.days(2).toMilliseconds();
>> private long currentMaxTimestamp;
>>
>> @Override
>> public long extractTimestamp(Tuple3<DateTime, String, String>
>> element, long previousElementTimestamp) {
>> long timestamp = element.f0.getMillis();
>> currentMaxTimestamp = Math.max(timestamp,
>> currentMaxTimestamp);
>> return timestamp;
>> }
>>
>> @Override
>> public Watermark getCurrentWatermark() {
>> return new Watermark(currentMaxTimestamp - maxOutOfOrderness);
>> }
>> }
>>
>> Thanks,
>> Chris
>>
>>
>>
>> --
>> View this message in context:
>> http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Multiple-windows-with-large-number-of-partitions-tp6521p6562.html
>> Sent from the Apache Flink User Mailing List archive. mailing list
>> archive at Nabble.com.
>>
>
>
> ------------------------------
> If you reply to this email, your message will be added to the discussion
> below:
>
> http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Multiple-windows-with-large-number-of-partitions-tp6521p6601.html
> To unsubscribe from Multiple windows with large number of partitions, click
> here
> <http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/template/NamlServlet.jtp?macro=unsubscribe_by_code&node=6521&code=Y2hyaXNAbmluamFtZXRyaWNzLmNvbXw2NTIxfC01MTI2ODMwNjU=>
> .
> NAML
> <http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>
--
View this message in context: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Multiple-windows-with-large-number-of-partitions-tp6521p6626.html
Sent from the Apache Flink User Mailing List archive. mailing list archive at Nabble.com.