Databases Reference
In-Depth Information
This schema mapping combines information in
Student
and
Enrolled
into
the
Takes
0
relation. Observe that
Takes
0
contains information about student ids,
name, and courses (as opposed to name, major, and course in
Takes
). Suppose I
consists of the following facts:
Student
.111;
John
;CS/;
Enrolled
.111;CS101/,
Student
.111;
John
;
Math
/;
Enrolled
.111;
MATH101
/.
The chase of I with
M
2
will produce the following target instance:
Takes
0
.111;
John
;
CS101
/,
Takes
0
.111;
John
;
MATH101
/.
The source facts
Student
.111;
John
;CS/and
Enrolled
.111;
CS
101/ together
trigger the s-t tgd in
M
2
to produce
Takes
0
.111;
John
;
CS101
/. In addition, the
source facts
Student
.111;
John
;
Math
/ and
Enrolled
.111;
MATH101
/ trigger
the s-t tgd in
M
2
to produce
Takes
0
.111;
John
;
MATH101
/ in the target. After
this, even though the source facts
Student
.111;
John
;
CS
/ and
Enrolled
.111,
MATH101
/ also trigger the s-t tgd in
M
2
, this chase step is not taken since the target
fact
Takes
.111,
John
,
MATH101
/ already exists in the target instance. It is easy
to observe that no other source facts would trigger the s-t tgd in
M
2
, and hence J
is the result of the chase. Also note that, as opposed to the previous example, there
is no need to generate nulls in the target, since
M
2
has no existentially quantified
variables (i.e., it is a GAV mapping).
3
An Ideal Scenario of Evolution
We start our exposition of the application of composition and inversion to schema
evolution, by considering first a relatively “simple” example of schema evolution.
For this section, we will refer to the schema evolution scenario that is graphically
illustrated in Fig.
7.2
.
We first assume the existence of a schema mapping
from a source schema
S
,
consisting of one relation
Takes
,toatargetschema
T
, consisting of two relations
Student
and
Enrolled
.The
Takes
relation contains tuples relating student ids
M
Schema S
Schema T
″
Schema S
Schema T
′
Student
Takes
″
Takes
sid
Takes
′
sid
major
major
sid
sid
cid
major
major
M
M
¢
M
¢¢
course
course
grade
Enrolled
sid
Course
cid
course
course
Fig. 7.2
Our first scenario of schema evolution
