Script expressions can use a variety of scripting languages. Currently supported languages are:
(XPath version 2 was supported until midPoint 3.5, but it is currently not supported)
Script Expression Structure
All the script expressions have the same internal structure no matter where and how they are used. The structure is illustrated in the following code snippet.
Meaning of individual script expression fields is as follows:
Language URL. Specifies the language in which the expression is written. If not specified it defaults to Groovy language.
Explicitly trace the execution of this expression. The execution will be logged on INFO level if this property is set to true to make sure that it will be visible in the logs.
The type of the expression result, either "list" or "scalar". If not set it will be determined automatically from the target data type for the expression. This setting should be used only if the automatic mechanism does not work as expected.
Values: relative (default) or absolute. See below for more details. If not specified and allowed by the expression usage it defaults to relative mode.
If set to true (which is the default) the script will be evaluated with null value as argument when the input is changed from or to an empty value. This generally works well for most cases. It may be set to false as an optimization.
Expression code. Depends on the script language used for the expression. This is usually a string, but it also may be an XML. Please note that the code is embedded in the XML therefore proper XML escaping is required
The XPath expressions used in midPoint are most frequently using variables. For example, the following expression will evaluate to the content of the
See Expression page for more generic information about the use of variables in expressions.
See MidPoint Script Library for more details.
Absolute and Relative Script Expressions
MidPoint always works with relative changes. MidPoint expressions are built for this mode of operation. The expression will receive every individual value on input and it has to transform that to output. MidPoint will take care of the logic around it. If the value was removed in the input, the result will be delete delta. If the value was added in the input the result will be add delta. MidPoint takes care of all that. The administrator only needs to specify an expression that transforms one value.
However, in some cases it might be useful to handle all values at once in an absolute way. Imagine, that you have multi-value UID attribute in LDAP and you want to select the right value based on DN. Then you need list of all the values. On the other hand, in relative approach, each value in UID attribute would be processed individually as a String.
The input to this script is a list of all input values. The output is a list of new values. MidPoint always operates in relative mode, therefore at the end of the evaluation midPoint will diff the value to create a delta. This does not change anything about that. However this mode of expression operation can be an advantage if you need to process all values as a group instead of processing every value one by one. E.g. in case that you want to choose a particular value or your algorithm depends on other values in some way.
The expressions are designed to be extensible and the expression language is not fixed. New expression languages may come in the future if there is a demand for them.