What are column datatypes in the Engine?

This article summarizes the different column datatypes in the AI & Analytics Engine.

In a dataset, values in the same column must be of the same datatype. The Engine currently supports the following datatypes:

Datatype Description Corresponding Type Examples
Numeric Whole numbers (Integers) and fractional numbers (real)
  • Julia: Float32, Float64, Int, Int32, Int64, BigInt

  • Python: double, int

  • R: numeric, double, int, bigint

  • -8, -5, -3, 0, 1, 3, 5, 8

  • -8.53, -0.001, 1.358

Boolean True or False
  • Julia: Bool

  • Python: bool, numpy.bool_

  • R: logical

  • True, False
Text

Free-form text

Category labels

JSON strings of attribute-value pairs

JSON string representation of arrays/lists

  • Julia: String, CategoricalVector

  • Python: str, object (pandas)

  • R: character

  • John Doe

  • New York

  • sunny, cloudy, windy

  • {"shape": "circle", "color": "blue", "size": 5}

  • [1, 2, "a", true, null]

  • [{"height": 183, "weight": 80}, {"height": 159, "weight": 60}]

DateTime Timestamp indicating date and time
  • Julia: Date, DateTime

  • Python: str, object (pandas)

  • R: Date, POSIXt

  • 2015-04-03 12:34:56

  • 2015-04-03