Filter condition reference and expression language

The following topics provide in-depth information about full-text and attribute conditions and how to use the expression language.

Full-text syntax
The full-text editor supports full-text syntax.
Full-text macros
The IBM® StoredIQ® full-text filter supports a set of predefined macros.
Indexed annotations
To search for indexed annotations (terms indexed when a cartridge is applied), you can create filters with the query format ia:filter_term.
Regular expressions
To search for patterns, you can create full-text filters with regular expressions. The format for a regular expression query is re:"regex".
Extended ASCII characters
Extended ASCII characters are handled several different ways in the full-text filter. Some are special characters while others are treated as spaces. Some are treated as letters that depend on the spaces around the characters.
Attribute condition reference
The following section explains how to type in a reference to an attribute condition.
Specification of attribute conditions
Attributes, entity, tag, or query can be divided up in several ways.
Type of object to match
For any attribute condition, you can specify what object that condition must match.
Multiple matching values for operators
You can give multiple matching values that use the following operators.
At-least expressions
On occasion, it is helpful to specify that some number of conditions must be true before an object is considered to match. These expressions are called at-least expressions because at least X subexpressions must match an object before the object matches the overall expression.
Parent and child conditions
This parent and child conditions represent advanced function that must be undertaken only by experienced administrators.
No-inheritance conditions
The no-inheritance conditions represent advanced function that must be undertaken only by experienced administrators.
Sampling
Checking a small, randomly selected subset of an otherwise huge infoset for compliance or violation of documents provides quicker results than scanning the whole infoset. Create a statistically representative random subset of an infoset by modifying any filter with a SAMPLE expression.

Parent topic: Building a filter