-
Notifications
You must be signed in to change notification settings - Fork 24.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Boxplot Aggregation #51948
Add Boxplot Aggregation #51948
Changes from 1 commit
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,204 @@ | ||
[role="xpack"] | ||
[testenv="basic"] | ||
[[search-aggregations-metrics-boxplot-aggregation]] | ||
=== Boxplot Aggregation | ||
|
||
A `boxplot` metrics aggregation that computes boxplot of numeric values extracted from the aggregated documents. | ||
These values can be extracted either from specific numeric fields in the documents, or be generated by a provided script. | ||
|
||
The `boxplot` aggregation returns essential information for making a https://en.wikipedia.org/wiki/Box_plot[box plot]: minimum, maximum | ||
median, first quartile (25th percentile) and third quartile (75th percentile) values. | ||
|
||
==== Syntax | ||
|
||
A `boxplot` aggregation looks like this in isolation: | ||
|
||
[source,js] | ||
-------------------------------------------------- | ||
{ | ||
"boxplot": { | ||
"buckets_path": "my_cardinality_agg" | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Typo re: "buckets_path" instead of "field"? |
||
} | ||
} | ||
-------------------------------------------------- | ||
// NOTCONSOLE | ||
|
||
Let's look at a boxplot representing load time: | ||
|
||
[source,console] | ||
-------------------------------------------------- | ||
GET latency/_search | ||
{ | ||
"size": 0, | ||
"aggs" : { | ||
"load_time_boxplot" : { | ||
"boxplot" : { | ||
"field" : "load_time" <1> | ||
} | ||
} | ||
} | ||
} | ||
-------------------------------------------------- | ||
// TEST[setup:latency] | ||
<1> The field `load_time` must be a numeric field | ||
|
||
The response will look like this: | ||
|
||
[source,console-result] | ||
-------------------------------------------------- | ||
{ | ||
... | ||
|
||
"aggregations": { | ||
"load_time_boxplot": { | ||
"min": 0.0, | ||
"max": 990.0, | ||
"q1": 165.0, | ||
"q2": 445.0, | ||
"q3": 725.0 | ||
} | ||
} | ||
} | ||
-------------------------------------------------- | ||
// TESTRESPONSE[s/\.\.\./"took": $body.took,"timed_out": false,"_shards": $body._shards,"hits": $body.hits,/] | ||
|
||
As you can see, the aggregation will return a calculated value for each percentile | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Looks like copy pasta leftovers :) |
||
in the default range. If we assume response times are in milliseconds, it is | ||
immediately obvious that the webpage normally loads in 10-725ms, but occasionally | ||
spikes to 945-985ms. | ||
|
||
==== Script | ||
|
||
The boxplot metric supports scripting. For example, if our load times | ||
are in milliseconds but we want percentiles calculated in seconds, we could use | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. "percentiles" typo There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. hmm, I am not sure how to call them here. Technically they are percentiles :) Values? There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Whoops, yeah this is probably fine. Was thinking it was a copy/paste leftover. |
||
a script to convert them on-the-fly: | ||
|
||
[source,console] | ||
-------------------------------------------------- | ||
GET latency/_search | ||
{ | ||
"size": 0, | ||
"aggs" : { | ||
"load_time_boxplot" : { | ||
"boxplot" : { | ||
"script" : { | ||
"lang": "painless", | ||
"source": "doc['load_time'].value / params.timeUnit", <1> | ||
"params" : { | ||
"timeUnit" : 1000 <2> | ||
} | ||
} | ||
} | ||
} | ||
} | ||
} | ||
-------------------------------------------------- | ||
// TEST[setup:latency] | ||
|
||
<1> The `field` parameter is replaced with a `script` parameter, which uses the | ||
script to generate values which percentiles are calculated on | ||
<2> Scripting supports parameterized input just like any other script | ||
|
||
This will interpret the `script` parameter as an `inline` script with the `painless` script language and no script parameters. To use a | ||
stored script use the following syntax: | ||
|
||
[source,console] | ||
-------------------------------------------------- | ||
GET latency/_search | ||
{ | ||
"size": 0, | ||
"aggs" : { | ||
"load_time_boxplot" : { | ||
"boxplot" : { | ||
"script" : { | ||
"id": "my_script", | ||
"params": { | ||
"field": "load_time" | ||
} | ||
} | ||
} | ||
} | ||
} | ||
} | ||
-------------------------------------------------- | ||
// TEST[setup:latency,stored_example_script] | ||
|
||
[[search-aggregations-metrics-boxplot-aggregation-approximation]] | ||
==== Boxplot values are (usually) approximate | ||
|
||
The algorithm used by the `boxplot` metric is called TDigest (introduced by | ||
Ted Dunning in | ||
https://github.com/tdunning/t-digest/blob/master/docs/t-digest-paper/histo.pdf[Computing Accurate Quantiles using T-Digests]). | ||
|
||
[WARNING] | ||
==== | ||
Boxplot as other percentile aggregations are also | ||
https://en.wikipedia.org/wiki/Nondeterministic_algorithm[non-deterministic]. | ||
This means you can get slightly different results using the same data. | ||
==== | ||
|
||
[[search-aggregations-metrics-boxplot-aggregation-compression]] | ||
==== Compression | ||
|
||
Approximate algorithms must balance memory utilization with estimation accuracy. | ||
This balance can be controlled using a `compression` parameter: | ||
|
||
[source,console] | ||
-------------------------------------------------- | ||
GET latency/_search | ||
{ | ||
"size": 0, | ||
"aggs" : { | ||
"load_time_boxplot" : { | ||
"boxplot" : { | ||
"field" : "load_time", | ||
"compression" : 200 <1> | ||
} | ||
} | ||
} | ||
} | ||
-------------------------------------------------- | ||
// TEST[setup:latency] | ||
|
||
<1> Compression controls memory usage and approximation error | ||
|
||
The TDigest algorithm uses a number of "nodes" to approximate percentiles -- the | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. I'm assuming this is the same as the Percentiles page? Maybe we should link to it instead? Or maybe there's a fancy way to embed a snippet from a different page? @nik9000 do you happen to know? There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. |
||
more nodes available, the higher the accuracy (and large memory footprint) proportional | ||
to the volume of data. The `compression` parameter limits the maximum number of | ||
nodes to `20 * compression`. | ||
|
||
Therefore, by increasing the compression value, you can increase the accuracy of | ||
your percentiles at the cost of more memory. Larger compression values also | ||
make the algorithm slower since the underlying tree data structure grows in size, | ||
resulting in more expensive operations. The default compression value is | ||
`100`. | ||
|
||
A "node" uses roughly 32 bytes of memory, so under worst-case scenarios (large amount | ||
of data which arrives sorted and in-order) the default settings will produce a | ||
TDigest roughly 64KB in size. In practice data tends to be more random and | ||
the TDigest will use less memory. | ||
|
||
==== Missing value | ||
|
||
The `missing` parameter defines how documents that are missing a value should be treated. | ||
By default they will be ignored but it is also possible to treat them as if they | ||
had a value. | ||
|
||
[source,console] | ||
-------------------------------------------------- | ||
GET latency/_search | ||
{ | ||
"size": 0, | ||
"aggs" : { | ||
"grade_boxplot" : { | ||
"boxplot" : { | ||
"field" : "grade", | ||
"missing": 10 <1> | ||
} | ||
} | ||
} | ||
} | ||
-------------------------------------------------- | ||
// TEST[setup:latency] | ||
|
||
<1> Documents without a value in the `grade` field will fall into the same bucket as documents that have the value `10`. |
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -378,7 +378,7 @@ public void testGetProperty() throws IOException { | |
iw.addDocument(singleton(new NumericDocValuesField("number", 7))); | ||
iw.addDocument(singleton(new NumericDocValuesField("number", 1))); | ||
}, (Consumer<InternalGlobal>) global -> { | ||
assertEquals(1.0, global.getDocCount(), 2); | ||
assertEquals(2, global.getDocCount()); | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. 😅 |
||
assertTrue(AggregationInspectionHelper.hasValue(global)); | ||
assertNotNull(global.getAggregations().asMap().get("min")); | ||
|
||
|
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,63 @@ | ||
/* | ||
* Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one | ||
* or more contributor license agreements. Licensed under the Elastic License; | ||
* you may not use this file except in compliance with the Elastic License. | ||
*/ | ||
|
||
package org.elasticsearch.xpack.analytics.boxplot; | ||
|
||
import org.elasticsearch.search.aggregations.metrics.NumericMetricsAggregation; | ||
|
||
public interface Boxplot extends NumericMetricsAggregation.MultiValue { | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Do we need these any more? There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Hmm, I'm not sure tbh. We might want to keep them around as a bridge to the HLRC? E.g. Or maybe useless bloat now that the TC is gone and we can get rid of them... but probably better to investigate that in a separate PR and apply it to all aggs if we decide we can remove them? There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. The HLRC won't have There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Argh, you're right. Keep forgetting this is in xpack. 👍 probably not necessary given the current situation |
||
|
||
/** | ||
* @return The minimum value of all aggregated values. | ||
*/ | ||
double getMin(); | ||
|
||
/** | ||
* @return The maximum value of all aggregated values. | ||
*/ | ||
double getMax(); | ||
|
||
/** | ||
* @return The first quartile of all aggregated values. | ||
*/ | ||
double getQ1(); | ||
|
||
/** | ||
* @return The second quartile of all aggregated values. | ||
*/ | ||
double getQ2(); | ||
|
||
/** | ||
* @return The third quartile of all aggregated values. | ||
*/ | ||
double getQ3(); | ||
|
||
/** | ||
* @return The minimum value of all aggregated values as a String. | ||
*/ | ||
String getMinAsString(); | ||
|
||
/** | ||
* @return The maximum value of all aggregated values as a String. | ||
*/ | ||
String getMaxAsString(); | ||
|
||
/** | ||
* @return The first quartile of all aggregated values as a String. | ||
*/ | ||
String getQ1AsString(); | ||
|
||
/** | ||
* @return The second quartile of all aggregated values as a String. | ||
*/ | ||
String getQ2AsString(); | ||
|
||
/** | ||
* @return The third quartile of all aggregated values as a String. | ||
*/ | ||
String getQ3AsString(); | ||
|
||
} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I wonder, should we snake_case this to
box_plot
? I personally don't really like snake casing but it would be more consistent (date_histogram
,significant_terms
, etc).There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Wikipedia states both spellings and
boxplot
was shorter, but I don't have strong feelings about it.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In that case, boxplot it is. I like shorter too