Github user aljoscha commented on a diff in the pull request:
https://github.com/apache/flink/pull/3191#discussion_r97828763
--- Diff: docs/dev/windows.md ---
@@ -622,133 +690,138 @@ input
</div>
</div>
-## Dealing with Late Data
+## Triggers
-When working with event-time windowing it can happen that elements arrive
late, i.e the
-watermark that Flink uses to keep track of the progress of event-time is
already past the
-end timestamp of a window to which an element belongs. Please
-see [event time](./event_time.html) and especially
-[late elements](./event_time.html#late-elements) for a more thorough
discussion of
-how Flink deals with event time.
+A `Trigger` determines when a window (as formed by the `WindowAssigner`)
is ready to be
+processed by the *window function*. Each `WindowAssigner` comes with a
default `Trigger`.
+If the default trigger does not fit your needs, you can specify a custom
trigger using `trigger(...)`.
-You can specify how a windowed transformation should deal with late
elements and how much lateness
-is allowed. The parameter for this is called *allowed lateness*. This
specifies by how much time
-elements can be late. Elements that arrive within the allowed lateness are
still put into windows
-and are considered when computing window results. If elements arrive after
the allowed lateness they
-will be dropped. Flink will also make sure that any state held by the
windowing operation is garbage
-collected once the watermark passes the end of a window plus the allowed
lateness.
+The trigger interface provides five methods that react to different
events:
-<span class="label label-info">Default</span> By default, the allowed
lateness is set to
-`0`. That is, elements that arrive behind the watermark will be dropped.
+* The `onElement()` method is called for each element that is added to a
window.
+* The `onEventTime()` method is called when a registered event-time timer
fires.
+* The `onProcessingTime()` method is called when a registered
processing-time timer fires.
+* The `onMerge()` method is relevant for stateful triggers and merges the
states of two triggers when their corresponding windows merge, *e.g.* when
using session windows.
+* Finally the `clear()` method performs any action needed upon removal of
the corresponding window.
-You can specify an allowed lateness like this:
+Any of these methods can be used to register processing- or event-time
timers for future actions.
-<div class="codetabs" markdown="1">
-<div data-lang="java" markdown="1">
-{% highlight java %}
-DataStream<T> input = ...;
+### Fire and Purge
-input
- .keyBy(<key selector>)
- .window(<window assigner>)
- .allowedLateness(<time>)
- .<windowed transformation>(<window function>);
-{% endhighlight %}
-</div>
+Once a trigger determines that a window is ready for processing, it fires.
This is the signal for the window operator to emit the result of the current
window. Given a window with a `WindowFunction`
+all elements are passed to the `WindowFunction` (possibly after passing
them to an evictor).
+Windows with `ReduceFunction` of `FoldFunction` simply emit their eagerly
aggregated result.
-<div data-lang="scala" markdown="1">
-{% highlight scala %}
-val input: DataStream[T] = ...
+When a trigger fires, it can either `FIRE` or `FIRE_AND_PURGE`. While
`FIRE` keeps the contents of the window, `FIRE_AND_PURGE` removes its content.
+By default, the pre-implemented triggers simply `FIRE` without purging the
window state.
-input
- .keyBy(<key selector>)
- .window(<window assigner>)
- .allowedLateness(<time>)
- .<windowed transformation>(<window function>)
-{% endhighlight %}
-</div>
-</div>
+<span class="label label-danger">Attention</span> When purging, only the
contents of the window are cleared. The window itself is not removed and
accepts new elements.
--- End diff --
This is a bit tricky because for non-merging windows there is nothing that
could be removed except the elements. Maybe write that PURGING will simply
remove the contents of the window and will leave any eventual meta information
intact and will also leave the Trigger state intact.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---