Overview

Attribute functions are used for performing calculations on data in PhixFlow. The most common uses are:

• populating output attributes;
• building filters.

This page explains the concepts you need to understand how functions work in PhixFlow. The concepts are should be fairly straightforward for anyone used to using Excel, or any type of database. However, it is important to understand them before writing the complex transformations that you can achieve using functions in PhixFlow.

Terminology and Conventions

In examples in the following descriptions, functions are given as out.calls = ifNull(in.numCalls, 0, 1). In this format, there are three sections of each expression:

• The first part of the example - out.calls - is either the output attribute name or an input attribute name (for an input filter).
• The middle part of the example - = - is the operator that links the two main parts of the expression. In a filter this can take a number of value, e.g. >, >=, contains. For an output attribute function this is always =, that is, that output attribute is set to the resulting value of the function on the right hand side.
• The final part of the example - ifNull(in.numCalls, 0, 1) – is the function itself.

Where a function can take, as one of its arguments, a list of values this is expressed in square brackets, e.g.:
function([value1, value2, ...])
See Arrays and Record Sets for details of how lists are handled in functions. Such a list is often called a vector. Indexes: always start at 1. Many functions return an index, indicating the position in a list at which a matched value has been found. E.g. see getElement.

Data Types

Functions can take the following data types:

• float: a floating point number
• integer: an integer
• string: a string
• date: there is no actual date type is PhixFlow; dates are handled as string, but these must have a particular format, either:
  • "yyyymmdd.hhmmss", or
  • "yyyymmdd".

Lists of Values

Where a function can take a list of values as one of its arguments, the argument will often be an attribute of an input table, or an expression based on an attribute of an input table. For example: sum(in.value) where in is an alias for an input table containing the numeric attribute value. When used as part of a merge or aggregate function, for example, the argument in.value will resolve to a list of values - all the values of value associated with a particular key combination.

Expression Operators

The following relational operators are available:

• ! (not; reverses the value of a logical expression)
• && (and; can also use the attribute function and)
• || (or; can also use the attribute function or)
• > (greater than; can also use the attribute function gt)
• < (less that; can also use the attribute function lt)
• != (not equals; can also use the attribute function ne)
• == (equals; can also use the attribute function eq)
• >= (greater than or equal to; can also use the attribute function ge)
• <= (less than or equal to; can also use the attribute function le)

The following arithmetic operators are available:

• + (add; can also use the attribute function sum)
• - (minus; can also use the attribute function sum, reversing the sign of the second operand)
• / (divide)
• * (multiply)

Responsive Functions 

To improve performance, a sub-set of functions, listed below, can be evaluated by the browser (client) rather than the server. These are useful where you want an expression to be evaluated quickly, for example in formatting rules.

If an expression includes:

• only the following, client-calculated functions, the whole expression is evaluated on the client.
• any other function, the expression is evaluated on the server. 

Functions that can be evaluated on the client: