Skip to main content

bin Search Operator

The bin operator assigns output results to user defined bins. A bin is configured to hold a range of values that can be used for sorting results in a histogram and further aggregation. It is a quick and effective way to visualize the distribution of data.

Syntax

bin <numeric field> width=<#>[, min=<#>][, max=<#>]

Where:

  • <numeric field> is the field you want to assign to bins. The data type needs to be numerical. Required.
  • <width> is the bin width as a floating point number. Required.
  • <min> is the lower boundary of the results as a floating point number. All data points with values less than the min are assigned to the first bin. Optional.
  • <max> is the upper boundary of the results as a floating point number. All data points with values more than the max are assigned to the last bin. Optional.

Output Fields

  • _bin_label is the default alias field, which has a standard interval representation.
  • _bin_lower is the lower boundary of the bin interval.
  • _bin_upper is the upper boundary of the bin interval.
tip

You can sort by the lower or upper boundary of the bin interval.

Examples

Distribution of error counts

A query counting error messages and using the bin operator to see the distribution of error counts based on bins with a width of 100:

_sourceCategory=stream error
| timeslice 1m
| count by _timeslice
| bin _count width=100.0
| count by _bin_label, _bin_lower
| sort by _bin_lower

bin operator charted.png

Latency distribution

A query parsing the latency of a function call and using the bin operator to see latency distribution over time:

_sourceCategory=analytics
| parse "ms: *" as time
| bin time width=10.0, min = 0.0, max = 500.0

You can aggregate bins further:

_sourceCategory=analytics
| parse "ms: *" as time
| bin time width=10, min = 0, max = 500
| count by _bin_label, _bin_upper
| sort by _bin_upper

example bin.png

Sumo Logic YouTubeSumo Logic Twitter
Legal
Privacy Statement
Terms of Use

Copyright © 2022 by Sumo Logic, Inc.