Database Reference
In-Depth Information
Fragmentation Transparency
A DDBMS supports data fragmentation if the DDBMS can divide and manage a logical object, such as the
records in a table, among the various locations under its control. The main purpose of data fragmentation is
to place data at the location where the data is most often accessed.
Suppose Premiere Products has a local DBMS at each of its three warehouses and wants to fragment its
Part table data, which is shown in Figure 9-2, by placing the data for the parts stored in a warehouse in that
warehouse
'
s local database.
Part
PartNum
Description
OnHand
Class
Warehouse
Price
AT94
Iron
50
HW
3
$24.95
280
BV06
Home Gym
45
SG
2
$794.95
CD52
Microwave Oven
32
AP
1
$165.00
DL71
Cordless Drill
21
HW
3
$129.95
DR93
Gas Range
8
AP
2
$495.00
DW11
Washer
12
AP
3
$399.99
FD21
Stand Mixer
2
1
2
W
3
$159.95
KL62
Dryer
2
AP
1
$349.95
KT03
Dishwasher
8
AP
3
$595.00
KV29
Treadmill
9
SG
2
$1,390.00
FIGURE 9-2
Premiere Products Part table data
Using SQL-type statements, you can define the following fragments:
DEFINE FRAGMENT Part1 AS
SELECT PartNum, Description, OnHand, Class, Warehouse, Price
FROM Part
WHERE Warehouse
¼
1
DEFINE FRAGMENT Part2 AS
SELECT PartNum, Description, OnHand, Class, Warehouse, Price
FROM Part
WHERE Warehouse
¼
2
DEFINE FRAGMENT Part3 AS
SELECT PartNum, Description, OnHand, Class, Warehouse, Price
FROM Part
WHERE Warehouse
¼
3
Each fragment definition indicates which Part table data to select for the fragment. Note that the Part
table does not actually exist in any one place. Rather, the Part table exists in three pieces. You assign these
pieces, or fragments, to the databases located at the warehouses, as shown in Figure 9-3.
