Scale Example 1: ScaleMethod ('midrange') - Teradata Vantage

This example scales (normalizes) input data using the midrange method and the default values for the arguments Intercept and Multiplier (0 and 1, respectively).

Input

• input: scale_housing, which has data about houses

  The types column has categorical data; the other columns have numerical data. The id column identifies the rows. The table has some NULL values.

• statistic: created by calling the ScaleMap function inside the Scale call

SQL Call

SELECT * FROM Scale (
  ON ScaleMap (
    ON scale_housing
    USING
    TargetColumns ('[2:6]')
    MissValue ('omit')
  ) AS statistic DIMENSION
  ON scale_housing AS "input" PARTITION BY ANY
  USING
  ScaleMethod ('midrange')
  Accumulate ('id')
) AS dt ORDER BY id, price, lotsize;

Output

The output table has the midrange-scaled values for the input data set. As explained in the descriptions of the arguments Multiplier and Intercept, this is the formula for computing the scaled value X' from the input value X:

X' = intercept + multiplier * (X - location)/scale

These are the formulas for computing location and scale for the midrange method (from Scale Arguments):

• location = (maxX+minX)/2
• scale = (maxX-minX)/2

The values minX and maxX are the minimum and maximum values of X, respectively.

For example, consider row 1 of the price column in the input table and the following output table: