Hi everyone,
I am not 100% sure about this one, so I thought that I could set my
thoughts straight via the mailing list.
Here's the use case. You coGroup a data set of vertices with a data set of
edges. That gives you a complexity of* O(|V| * |E|)*, where |V| is the
total number of vertices and |
Were you able to resolve the issue?
I first thought its an JVM-version incompatibility, but this thread
suggests a serialization issue:
http://stackoverflow.com/questions/851491/java-deserialization-of-java-lang-integer-exception
By the way: your emails to the dev@ list sometimes end up in the "S
Hi Andra,
CoGroup is not implemented as a Cartesian product, so O(V*E) is not a very
accurate approximation.
All this depends on what you count. Let's assume single-node execution and
that everything fits in memory, and let's count comparisons and UDFs on
groups.
Then, coGroup sorts both inputs,
Hmmm, this seams like the stream is read differently than it is written.
Maybe one of the involved classes overrides the readObject() and
writeObject() methods, and implements them inconsistently. And you see if
that is the case in one of your data types?
On Wed, Jul 22, 2015 at 11:02 AM, Robert
Matthias J. Sax created FLINK-2392:
--
Summary: Instable test in flink-yarn-tests
Key: FLINK-2392
URL: https://issues.apache.org/jira/browse/FLINK-2392
Project: Flink
Issue Type: Bug
Stephan Ewen created FLINK-2393:
---
Summary: Add a "stateless at-least-once" mode for streaming
Key: FLINK-2393
URL: https://issues.apache.org/jira/browse/FLINK-2393
Project: Flink
Issue Type: Ne
Stefano Bortoli created FLINK-2394:
--
Summary: HadoopOutFormat OutputCommitter is default to
FileOutputCommiter
Key: FLINK-2394
URL: https://issues.apache.org/jira/browse/FLINK-2394
Project: Flink
Hi All,
I added a new document along the existing design documents in the wiki:
https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=60624830.
It's very bare-bones right now but we can flesh it out together. I put some
requirements that I think are good but please add other stuff or wri
Hi all!
Yesterday, we merged the framework for the new Flink runtime monitor
dashboard. New stub, based on angular.js and coffee script.
https://github.com/apache/flink/tree/master/flink-runtime-web
Anyone who is savvy with these technologies and wants to work on that part,
please ping! We'd be
Finally, Flink is a JavaScript project :-)
On Wed, Jul 22, 2015 at 6:22 PM, Stephan Ewen wrote:
> Hi all!
>
> Yesterday, we merged the framework for the new Flink runtime monitor
> dashboard. New stub, based on angular.js and coffee script.
>
> https://github.com/apache/flink/tree/master/flink-r
Hi
Once DataStream API is out of beta and both DataStream and DataSet API are
stabilized, will we be seeing Flink 1.0 release?
On April 28, 2014, Fabian Hueske personal guess of Flink 1.0 release in his
info.com interview was April 2016? Reference:
http://www.infoq.com/news/2015/04/hueske-apac
Haha!
On 07/22/2015 07:02 PM, Kostas Tzoumas wrote:
> Finally, Flink is a JavaScript project :-)
>
> On Wed, Jul 22, 2015 at 6:22 PM, Stephan Ewen wrote:
>
>> Hi all!
>>
>> Yesterday, we merged the framework for the new Flink runtime monitor
>> dashboard. New stub, based on angular.js and coffe
12 matches
Mail list logo