[
https://issues.apache.org/jira/browse/FLINK-1534?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14360034#comment-14360034
]
Akshay Dixit commented on FLINK-1534:
-------------------------------------
Hi Marton,
To get familiar with the codebase and flink-streaming, I've attempted to cobble
together a PR for FLINK-1450.
So during this, I came across various functions while adding a FoldFunction
that are implemented over the stream such as FlatMapFunction, FilterFunction
etc.
I was wondering why wouldn't a simple concatentation of FilterFunctions over
the stream not work for pattern matching. Because if the patterns to detect
would be like (eventA -> eventB), then MyDataStream.filter(x == eventA and x+1
== eventB) should work right?
Sorry if I'm completely off course here.
> GSoC project: Distributed pattern matching over Flink streaming
> ---------------------------------------------------------------
>
> Key: FLINK-1534
> URL: https://issues.apache.org/jira/browse/FLINK-1534
> Project: Flink
> Issue Type: New Feature
> Components: Streaming
> Reporter: Márton Balassi
> Priority: Minor
> Labels: gsoc2015, java, scala
>
> Pattern matching over streams is an important application. The general
> structure of a streaming pattern matching is the following:
> If A event follows B event then trigger some computation.
> The support for this feature is associated with complex event processing
> systems, however it is also adoptable for distributed setting, however it
> poses additional challenges.
> The Google Summer of Code student volunteering for this project is expected
> to have general knowledge of distributed systems and Java/Scala coding
> skills. The project includes research and implementation oriented taks.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
