Baked Fields

Function fields - also known as "BakedFields" - are inbuilt DataJet functions that can be applied to existing data fields in order rapidly create new fields.  Along with Expressions, DataSetFields and User-Defined Fields they play a key role in the transformation of data.

The engineering functions are organised into groups:

• General - general functions, including key creation, discrete value processing, access to indexes, joinlayer functions and quantiles.
• DateTimeTransform - transformations involving hours, minutes and seconds
• DateTransform - transformations involving years, months and days
• DMath - transformations producing decimal fields
• LMath - transformations that produce integer fields
• FLogic - logical functions that evaluate conditions and produce flag fields
• Geo - geographical functions, including distance calculations
• Statistical - distributions and transformation that support data modelling and cleaning
• String - transformations of string data
• UserAgent - extractions of web data

Creating Baked Fields

BakedFields (or functions) can be created in a variety of ways:

• From the Engineering | Functions menu
• From the Database Explorer | Right-click | Engineering | Functions options
• From Script
• By editing an existing Baked Field

Baked Fields are optimised functions that are designed to work quickly on very large datasets.   Not all functions are supported for all data and field types.  If a bakedfield function is not available for the particular field or data type that needs to be transformed, try one of the following:

• Expression Builder - (note, expressions may be slower than BakedFields)
• User-Defined Field
• Casting, or pre-engineering

See Engineering - Functions for more details.

JSON and Parameters

Baked Fields have a common parameter structure:

Example Baked field (YMDH - Year Month Date Hour)

{
  "method": "BuildBakedField",
  "project": "Election5",
  "targetTable": "PA_All",
  "overwrite": true,
  "name": "YMDH",
  "function": "datetimetransform",
  "p1": "YMDH(A)",
  "p2": "TimeStamp"
}

{
  "method": "BuildBakedField",
  "targetTable": "transactions",
  "overwrite": true,
  "name": "Engineering-DMath",
  "function": "dmath",
  "p1": "A/B",
  "p2": "profit",
  "p3": "price",
  "nullIsZero": true,
  "project": "RegressionDB2"
}