Database Reference
In-Depth Information
supported assignment
. The STP predicate yields true iff at least one supported
assignment is found. This completes our description of the STP predicate.
The
STP Algebra
in SECONDO defines other variants of the STP predi-
cate (e.g.,
stpatternexextendstream
). This operator is a triple
where
P
,
C
are the same as before, and
f
is an additional condition on the
time intervals of the supported assignments. One can express, for instance, that
the
Down
predicate in this query must be fulfilled for at least 2 minutes. The
stpatternexextendstream
is also a stream operator, not a predicate. It
extends every input tuple with attributes containing the time intervals on which
the pattern occurs. Since one trajectory might contain several matches of the
pattern, the
stpatternextendstream
copies the tuple, and extends every
copy with one match. The following example expresses the
stepwise descent
scenario:
P,C,f
1
...stpatternexextendstream[
2
Dive1: .SecondAltitudeDerivative < 0.0,
3
Lift: .SecondAltitudeDerivative >= 0.0,
4
Dive2: .SecondAltitudeDerivative < 0.0
;
5
stconstraint("Dive1", "Lift", vec("aa.bb")),
6
stconstraint("Lift", "Dive2", vec("aa.bb"));
7
(end("Lift") - start("Lift")) > OneMinute ]
8
filter[isdefined(.Dive1) and
9
(AverageDiveAngle(.Alt atperiods .Lift) < 30.0)]...
In this scenario, the aircraft alternates between dive and cruise during its final
approach. It is expressed as a sequence of increasing, decreasing, then again
increasing rate of descent. Line 7 asserts that the
Lift
event stays more than a
minute. Line 9 invokes the SECONDO function object
AverageDiveAngle
to
assert that the aircraft is flying almost horizontally during the
Lift
event, hav-
ing a slope of less than 30
◦
with the horizontal. The two queries in this section
finish in approximately 1 minute on the given data set with 17,851 trajecto-
ries (427,651 records). The SECONDO relation storing these flight trajectories
occupies approximately 172 MB of disk-space on a Linux 32 bit machine.
12.6.2 Exploring Patterns by Integrating MOD with Visual Analytics
So far, we have shown that the STP predicates and its variants are very flexible
and can be used to express arbitrarily complex patterns. In practice, tuning the
parameters of these operators is tricky. The integration with visual analytics
allows for fine tuning these parameters through user interaction. SECONDO and
V-Analytics realize such an integration scheme. They are integrated so that it
is possible to interchange query results in both directions. Typically the user
starts by loading the whole data set in the databases of the two systems. The
