apache / apache/beam

Unable to use ApproximateQuantiles.globally/ApproximateUnique.globally when inputs not windowed by GlobalWindows

Open
#20,301 0 comments 0 reactions 0 assignees View on GitHub
bug core java P3
Dominant language
Java
Stars
8.7k
Forks
4.7k
Avg merge
1d 20h
Merged PRs (30d)
196

Description

Unable to use ApproximateQuantiles.globally or ApproximateUnique.globally with input windowed not using GlobalWindows.
To make it run we need to set either
```

.withoutDefaults()

```

or
```

.asSingletonView()

```

Currently we can't call any of the above on ApproximateQuantiles.globally()/ApproximateUnique.globally as it does not return underlying Combine.globally, but PTransform or Globally in case of ApproximateUnique.

Example failing case:

```

PCollection elements = p.apply(GenerateSequence.from(0).to(100)
.withRate(1,Duration.millis(1)).withTimestampFn(Instant::new));

PCollection> input = elements
.apply(Window.into(SlidingWindows.of(Duration.millis(3)).every(Duration.millis(1))))

.apply(ApproximateQuantiles.globally(17));

```

It throws expected error from internal Combine.globally() transform:

```

Default values are not supported in Combine.globally() if the input PCollection is not windowed by
GlobalWindows. Instead, use Combine.globally().withoutDefaults() to output an empty PCollection if the
input PCollection is empty, or Combine.globally().asSingletonView() to get the default output of the
CombineFn if the input PCollection is empty.

```

Imported from Jira [BEAM-10005](https://issues.apache.org/jira/browse/BEAM-10005). Original Jira may contain additional context.
Reported by: darshanjani.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.