Calculates the percentile for each value relative to the rest of the data set.
Produces a percentile for each row of input. The rows are ranked by the value defined in the InputExpr property.
For this node, percentiles are defined as the percentage of all input values that fall below the current score. This definition parallels the Excel PERCENTRANK function.These percentiles are output in a separate field whose name is defined by the PercentileFieldName property. Percentiles are reported as floating point values without a percentage sign. In other words, for a value in the 20th percentile, its percentile field will contain the value ".2".
The input of this node either must be sorted by InputExpr before entering the node or must be sorted via the SortInput property.
If the reverse keyword is used to sort the input, then this node will produce a reverse percentile; in other words, the percent rank will show the percentage of all values above the current value, rather than below it.
Optionally specify whether to sort the input data. If true, sorts the input by InputExpr. If false, then the data is assumed to have been sorted already and is not sorted by this node. If Epsilon is provided, it is not used to determine identical values in sorting this data; rather, it is used only to determine identical values for calculating the percentiles.
The default value is False.
Specify which value will be used to determine the percentiles for the input data. This property could contain a simple field reference or a mathematical expression. This field cannot evaluate to a list or to a non-numeric value.
If the reverse keyword is used in this property, then the percent rank is changed to report the percentage of all input values that are above the current value rather than below it.
A value is required for this property.
Optionally specify a name for the output field for percentile values
The default value is "_Percentile".
Optionally specify a small, positive floating point value that defines the tolerance used to compare floating point numbers.
When comparing floating point numbers, it is common to use a tolerance to determine when two values are equal. This tolerance ensures that miniscule differences between two floating point values do not cause false negatives.
If InputExpr evaluates to a floating point value, then Epsilon cannot be blank. Otherwise, it is ignored.
Inputs and outputs