The SASE+ language has a high-level structure similar to SQL for ease of use. The design of the language, however, is centered on temporal event patterns that have not been sufficiently addressed in relational data processing. In particular, SASE+ has the following features:
- Sequencing: a sequence of events occur in specified order.
- Negation: an event pattern can address the non-occurrences of events.
- Kleene closure: a pattern can also be used to extract a finite yet unbounded number of events with a particular property.
- Parameterized predicates: different events are compared via value-based constraints as well as temporal constraints.
- Sliding windows: the entire pattern needs to occur within a specified period of time.
The overall structure of the SASE+ language is:
[FROM < input stream >]
PATTERN < pattern structure >
[WHERE < pattern matching condition >]
[WITHIN < sliding window >]
[HAVING] < pattern filtering condition >
[RETURN < output specification >]
The FROM clause provides the name of an input stream. If it is omitted, the query refers to a default system input.
The PATTERN, WHERE and WITHIN clauses form the pattern matching block. The PATTERN clause specifies the structure of a pattern to be matched against the input stream. The WHERE clause, if present, imposes value-based constraints on the events addressed by the pattern. The WITHIN clause further specifies a sliding window over the entire pattern.
The evaluation of PATTERN, WHERE, and WITHIN clauses results in a stream of pattern matches; each consists of a unique sequence of events used to match the pattern. The HAVING clause further filters each pattern match by applying predicates on the constituent events.
Finally, the RETURN clause transforms each pattern match into a result event for output.
Detailed examples drawn from various real-world scenarios are presented in the next page.Last updated: 07/16/14 11:35:03