Histogram Generator

About

A histogram generator is a column data generator that generates a value according to its chance factor.

This generator is used to generate data that follows a probability distribution where the factor is the probability for the value.

Example

Distribution over week day (Varchar)

Example if you want to simulate that people may work the week-end once every 10 weeks, you will define this bucket list

Columns:
  - name: bucket_map
    type: varchar
    data-supplier:
      type: histogram
      arguments:
        Buckets:
          Monday: 10
          Tuesday: 10
          Wednesday: 10
          Thursday: 10
          Friday: 10
          Saturday: 1
          Sunday: 1

Normal Distribution over Arrival Time

Example of normal distribution over time

kind: generator
spec:
  MaxRecordCount: 30
  Columns:
    - name: id
      type: integer
      comment: A id column to see easily the number of values generated
      data-supplier:
        type: sequence
    - name: bucket_map
      type: time
      comment: A column with a histogram generator that generates an uniform distribution of time
      data-supplier:
        type: histogram
        arguments:
          Buckets:
            "8:45": 0.05
            "8:50": 0.5
            "8:55": 0.22
            "9:00": 0.4
            "9:05": 0.22
            "9:10": 0.5
            "9:15": 0.05

Arguments

Buckets

This data-supplier has only one argument that defines the histogram namely, the buckets.

A bucket is a value and its chance factor

Data Type

The below data type are supported:

Name	Yaml Format	Example
Integer	d	8
Double	d.dd	8.00
Date	YYYY-MM-DD	1970-01-01
Timestamp	YYYY-MM-DD HH:MM:SS	1970-01-01 00:00:00
Time	"HH:MM", "HH:MM:SS", "HH:MM:SS.SSS"	"08:00" quote is mandatory
Varchar	".*"	"a text"

Why the time must be quoted? Yaml does not support time as a type. The time string should be quoted.

How to define a Bucket definition in a data set

Note that the data set and the entity generator creates histogram from resources.

The data being defined by the column attribute
the chance factor being defined by the probability, weight or factor column