Database Reference
In-Depth Information
Sequence Scan and Construction (SSC):
While sequence scan
and construction are technically different operators, they are al-
ways used together. The corresponding component is referred to
as
SSC
. When a query contains the
SEQ
construct in the language,
all the
positive
components of the
SEQ
specification are handled
by the sequence scan and construction operators. Thus, a sub-
sequence of the original specification is handled by this pair of
constructs. Thus, the function of the
SSC
is to transform a stream
of events into a stream of event sequences, each of which represents
a unique match of specified
SSC
sub-sequence type.
Selection:
This is the commonly used operator in relational query
processing. In this case, this operator is used to filter each event
sequence by applying different predicates.
Window:
The window operator imposes the constraint of the
WITHIN
clause. For each event sequence, it checks if the temporal
difference between the first and last events is less than the specified
window
T
.
Negation:
The negation operator handles the negative compo-
nents of a
SEQ
construct which have been ignored by SSC.
Transformation:
This operator converts each event sequence to
a composite event by concatenating attributes of all the events in
the sequence.
Another recent method for event processing with RFID data has been
proposed in [9]. This method has the ability to query different readers for
data in order to make key real time inferences for events. In addition,
methods have been designed to work with the code embedded in the
RFID tags for event processing. The EPC tags represents a string, in
which different portions of the string correspond to different parts of
the information about the product. Therefore, any algorithm needs to
be able to work effectively in terms of deciphering the importance of
different portions of the string for event processing. The approach in [9]
shows how to extend an SQL-based query language in order to make it
suitable for event processing in the context of RFID data. This can be an
advantage in many scenarios, because users are often more familiar with
SQL-like languages. Because of this, recent systems for event processing
[19] have generally tried to work with extensions of the SQL language
for event processing.
It has been observed [6] that stream event detection algorithms can be
generally formulated as
pattern matching algorithms
over data streams.
