Skip to main content
Version: v4.22

deduplicate

Removes duplicate events based on the values of one or more fields.

Synopsis

deduplicate [<extractor>...]
[--limit <count>] [--distance <count>] [--timeout <duration>]

Description

The deduplicate operator removes duplicates from a stream of events, based on the value of one or more fields.

You have three independent configuration options to customize the operator's behavior:

  1. Limit: the multiplicity of the events until they are supressed as duplicates. A limit of 1 is equivalent to emission of unique events. A limit of N means that events with a unique key (defined by the fields) get emitted at most N times. For example, GGGYBYYBGYGB with a limit of 2 yields GGYBYB.
  2. Distance: The number of events in sequence since the last occurrence of a unique event. For example, deduplicating a stream GGGYBYYBGYGB with distance 2 yields GYBBGYB.
  3. Timeout: The time that needs to pass until a surpressed event is no longer considered a duplicate. When an event with surpressed key is seen before the timeout is reached, the timer resets.

The diagram below illustrates these three options. The different colored boxes refer to events of different schemas.

<extractor>...

A comma-separated list of extractors that identify the fields used for deduplicating. Missing fields are treated as if they had the value null.

Defaults to the entire event.

--limit <count>

The number of duplicates allowed before they are removed.

Defaults to 1.

--distance <count>

Distance between two events that can be considered duplicates. Value of 1 means only adjacent events can be considered duplicates. 0 means infinity.

Defaults to infinity.

--timeout <duration>

The amount of time a specific value is remembered for deduplication. For each value, the timer is reset every time a match for that value is found.

Defaults to infinity.

Examples

Consider the following data:

{"foo": 1, "bar": "a"}
{"foo": 1, "bar": "a"}
{"foo": 1, "bar": "a"}
{"foo": 1, "bar": "b"}
{"foo": null, "bar": "b"}
{"bar": "b"}
{"foo": null, "bar": "b"}
{"foo": null, "bar": "b"}

For deduplicate --limit 1, all duplicate events are removed:

{"foo": 1, "bar": "a"}
{"foo": 1, "bar": "b"}
{"foo": null, "bar": "b"}
{"bar": "b"}

If deduplicate bar --limit 1 is used, only the field bar is considered when determining whether an event is a duplicate:

{"foo": 1, "bar": "a"}
{"foo": 1, "bar": "b"}

And for deduplicate foo --limit 1, only the field foo is considered. Note, how the missing foo field is treated as if it had the value null, i.e., it's not included in the output.

{"foo": 1, "bar": "a"}
{"foo": null, "bar": "b"}