Variable types in Process Designer

You can use the variable types provided by the system toolkits, such as the System Data toolkit, or you can create custom business objects, depending on the requirements of the business data included in your process.

During the Process Designer installation, system toolkits are imported into the Process Center repository so that each process application and toolkit that you create has access to a common system data. System toolkits provide the following categories of variables:

Base types: Base types allow you to create custom variable types called business objects. For a list of all the base types, see the following table.

System types: System types are provided variable types that cannot be modified. For a list of all the system types, see JavaScript API in Process Designer.

This table provides more information about the base type business objects provided in system toolkits.

Table 1. Provided base types
Base types	Description
`String`	Allows alpha-numeric characters to be entered into the variable
`Integer`	Accepts digits without a decimal place in the range from -2147483648 to 2147483647, such as `45` or `20`
`Decimal`	Accepts digits with decimal places with IEEE 754 double precision, such as `45.3` or `20.13`
`Date`	Allows date and time formats to be entered into the variable
`Time`	Allows date formats to be entered into the variable as times. The user enters a time; before the variable is entered into the symbol table, it is converted into a date.
`Selection`	Allows you to provide a list of possible entries to a user, from which the user can select only one. A selection is a list of different values; each value is a string. At run time, a selection variable appears on a coach form as a list or as radio buttons.
`Boolean`	Accepts either `true` or `false` as values. The Boolean value appears at run time on a coach form as a check box.
`Structure`	To use a structure type, you must create a custom structure type and define its properties. A structure regroups business data that is related to the same subject. For example, a `Customer` structure might contain elements such as `lastName`, `firstName`, `homeNumber`, `streetAddress`.

Custom variable types

Custom variable types are defined by using business objects. If the predefined business objects that are provided in the system toolkits do not represent your needs, you can create your own business objects.

Validation for custom variable types

You can define validation settings for custom variable types that use a simple type definition. These validation rules are checked in the following situations:

When you assign a value to a variable in a business process definition (BPD) or service of such a type, such as when a script assigns a value to a tw.local variable or an output mapping where a service result is mapped into a tw.local variable, the value is validated.
When you have a heritage human service with a heritage coach, the input fields that are bound to a service variable are validated when you submit the heritage coach. To enable validation on the coach view, you must define the validation rules again in the coach view configuration where you bound your variable.

Validation for Integer and Decimal custom variable types

For Integer and Decimal business object definitions, you can define minimum and maximum values to specify a range. For example, if you specify a minimum value of 0 and a maximum value of 100, both inclusive, you have a type that could be used for a variable where users specify a percentage that cannot be negative or more than 100%.

By specifying the precision and scale, you can restrict the decimal number and the decimal places of a number.

IBM BPM version 8570 cumulative fix 2016.06 You specify how the validation settings must be interpreted in the business object configuration of the server configuration.

Table 2. Precision and scale validation properties
Properties	Type	Default value	Description
`precision-validation-on-server-enabled`	Boolean	false	Used to enable the validation of the precision setting when a variable is set. If it is not enabled, it is checked only in heritage coaches.
`precision-validation-strip-trailing-zeros`	Boolean	false	Used to remove trailing zeros from behind the decimal point before the precision is validated
`precision-validation-type`	enum (equals, lessOrEquals)	equals	If this property is set to `equal`, the precision of the variable value must equal what is specified in the variable type. If the property is set to `lessOrEquals`, the precision of the variable value must equal or be less than what is specified in the variable type.
`scale-validation-strip-trailing-zeros`	Boolean	false	Used to remove trailing zeros from behind the decimal point before the scale is validated
`scale-validation-type`	enum (equals, lessOrEquals)	equals	If this property is set to `equals`, the scale of the variable value must equal what is specified in the variable type. If the property is set to `lessOrEquals`, the scale of the value must equal or be less than what is specified in the variable type.

The following example shows how you can set the precision value as 5 and scale value as 2 for a variable type.

If you set the precision-validation-on-server-enabled property to true but keep all other settings at their default, the following value is valid:

123.45

The following values are invalid:

1234.56 (wrong precision, 6 instead of 5)
12.34 (wrong precision, 4 instead of 5)
12.345 (wrong scale, 3 instead of 2)
1234.1 (wrong scale, 1 instead of 2)

The following JavaScript expression also causes a validation error because JavaScript removes trailing zeros at run time before it assigns the value:

tw.local.variable = 123.40

The following JavaScript expression does not cause a validation error because the string is validated correctly and then is automatically converted into a number to set the variable value:

tw.local.variable = "123.40"

To use the precision and scale settings as their XML schema equivalents totalDigits and fractionDigits, you must set the configuration flags as shown in the following table:

Table 3. Precision and scale validation settings equivalent to totalDigits and fractionDigits XML schema classes
Setting	Type
`precision-validation-on-server-enabled`	true
`precision-validation-strip-trailing-zeros`	true
`precision-validation-type`	lessOrEquals
`scale-validation-strip-trailing-zeros`	true
`scale-validation-type`	lessOrEquals

The same variable type that is used in the preceding sample then considers the following values valid:

123.45
12.34
123.4
123

The following values are still invalid:

1234.56 (wrong precision, 6 instead of a maximum of 5)
12.345 (wrong scale, 3 instead of a maximum of 2)

Numeric calculations in scripts

IBM® Business Process Manager uses JavaScript so that you can program within a process or service step. JavaScript is based on the ECMAScript specification that specifies double precision must be used for numbers. IBM BPM also uses double precision to store variable values.

When you perform calculations in scripts, be careful of limits that can come to this precision. For example, when you add 1000.06 + 0.01, the mathematically correct result is 1000.07. However, if you add the numbers in JavaScript, the result is 1000.0699999999999.

If you require mathematically correct calculations, perform these calculations outside of the JavaScript code. For example, you can use the BigDecimal and BigInteger classes to do calculations with full precision in a Java program. If you must store the result with full precision in a process or service variable, you can format the number as a string and use the String data type for your variable.

Restriction: A value of NaN or Infinity cannot be stored in the database for a process variable that is visible in Process Portal. If you calculate the value of such a variable, ensure that the result is not NaN or Infinity; otherwise, an error is written to the SystemOut.log file.