flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From fhueske <...@git.apache.org>
Subject [GitHub] flink pull request #2368: [FLINK-3899] Document window processing with Reduc...
Date Wed, 24 Aug 2016 07:44:50 GMT
Github user fhueske commented on a diff in the pull request:

    https://github.com/apache/flink/pull/2368#discussion_r76008397
  
    --- Diff: docs/apis/streaming/windows.md ---
    @@ -459,42 +459,106 @@ ready for processing. This allows to get the benefit of incremental
window compu
     the additional meta information that writing a `WindowFunction` provides.
     
     This is an example that shows how incremental aggregation functions can be combined with
    -a `WindowFunction`.
    +a `WindowFunction`.  The `FoldFunction`/`WindowFunction` example shows how to extract
the
    +ending event-time of a window of sensor readings that contain a timestamp, 
    +and the `ReduceFunction`/`WindowFunctions` example shows how to do eager window
    +aggregation (only a single element is kept in the window).
     
     <div class="codetabs" markdown="1">
     <div data-lang="java" markdown="1">
     {% highlight java %}
    -DataStream<Tuple2<String, Long>> input = ...;
    +DataStream<SensorReading> input = ...;
     
     // for folding incremental computation
     input
         .keyBy(<key selector>)
         .window(<window assigner>)
    -    .apply(<initial value>, new MyFoldFunction(), new MyWindowFunction());
    +    .apply(Long.MIN_VALUE, new MyFoldFunction(), new MyWindowFunction());
    +
    +/* ... */
    +
    +private static  class myFoldFunction implements FoldFunction<SensorReading, Long>
{
    +
    +    public Long fold(Long acc, SensorReading s) {
    +        return Math.max(acc, s.timestamp());
    +    }
    +}
    +
    +private static class MyWindowFunction implements WindowFunction<Long, Long, String,
TimeWindow> {
    +
    +    public void apply(String key, TimeWindow window, Iterable<Long> timestamps,
Collector<Long> out) {
    +            out.collect(timestamps.iterator().next());
    --- End diff --
    
    The example looks good, thanks! Two minor suggestions: 1) I think we can omit setting
key and time in the `FoldFunction`, 2) the `WindowFunction` could fetch the count in a separate
variable. This would make the `out.collect` line a bit shorter.
    
    Regarding the type restriction: You discovered a bug that we would like to fix but can't
until Flink 2.0.0 because we promoted the interface to be `@Public` and the API is stable
in Flink 1.0 releases. IMO it makes sense to point out this accidental restriction in the
documentation.



---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

Mime
View raw message