Database Reference
In-Depth Information
suburb
(su.)} are sets of linguistic labels defined respectively on attributes Price, Size and Location.
Figure 17 shows the summary hierarchy
H
R
provided by SAINTETIQ performed on
R
.
3.2.2 Expression and Evaluation of Queries
The querying mechanism proposed here intends to evaluate queries such as
Q
1
:
Example 3.1
(
Q
1
)
SELECT *
FROM
R
WHERE Location IN {
suburb
}
AND Size IN {
medium
OR
large
}
In the following, we denote by
X
the set of attributes specified in the user's input query
Q
and by
Y
the set of missing attributes which is the complement of
X
relatively to
R
:
XY=R
and
XY=φ
. Further, for
each attribute
AX
,
Q
A
denotes the set of required features defined on attribute
A
. For instance, given
Q
1
,
we have
X
={Location, Size}, Y={Price},
Q
Location
={
suburb
},
Q
Size
={
medium
,
large
}.
Let
z.A
be the set of descriptors that appear in a summary
z
on attribute
A
. When matching
z
with
a query
Q
to decide whether it corresponds to that query and can then be considered as a result, three
cases might occur:
a:
no correspondence (i.e.,
∧
A
∧
X,z.A
∩
Q
A
=
φ
). For one attribute or more,
z
has no required feature,
i.e., it shows none of the linguistic labels mentioned in query
Q
.
b:
exact correspondence (i.e.,
∧
A
∧
X,z.A
∧
Q
A
). The summary
z
matches the query
Q
semantics. It is
considered as a result.
Figure 17. Summary hierarchy H
R
of the data set R
