|
Category: ASE: Data Storage
Overflow pages
Overflow pages are created in
nonunique clustered indexes
on
APL tables
when a new row is inserted which has the same key value as the last row on an already full
data page.
A new data page will be assigned to the respective
page chain
and linked in. The new row will be stored on this new, so-called
overflow page.
The overflow page only stores rows with identical key values.
Sometimes there can be numerous identical values, if data is inserted into a column with a
nonunique index.
Example of an overflow page
Before the insert:
|
page # 1132
|
|
green
|
|
violet
|
|
yellow
|
|
blue
|
insert table_name ('color') values ('blue')
After the insert:
|
page # 1132
|
|
green
|
|
violet
|
|
yellow
|
|
blue
|
|
(overflow) page # 1158
|
|
blue
|
|
|
|
|
|
|
Characteristically, overflow pages are not directly referenced by the index pointers of
clustered indexes
Instead, the
next page pointer
is used to follow a chain of overflow pages. The next page pointer
is followed until the search value cannot be matched anymore.
|