Glossary - Relational Database Index Design and the Optimizers - page 298

SQL Server term table lookup for an access path that uses an index, but also reads
table rows.
CPU cache The CPU chip's high-speed memory used to store the most frequently
used program instructions and data.
Cursor A construction in embedded SQL for moving the result table into the
application program one row at a time with FETCH calls.

Database Logically related data; a collection of tables—DBMS dependent.
Data-partitioned secondary index Partitioning very large indexes, one index
partition per table partition, in order to reduce unavailability.
Data block prefetching An Oracle term used to represent the process whereby the
pointers are collected from an index slice so that multiple random I/Os can be started
to read the table rows in parallel.
DBMS, database management system The software used to provide, manage, and
control all aspects of the use of databases. Network and hierarchical systems have
now almost entirely been superseded by relational systems.
Data warehouse A business environment that provides consistent and time-dependent
data to support decision-making processes. Designed to support a large variety of
unpredictable queries and reports. It is often loaded from multiple operational systems
converted into a consistent format. It enables trend analysis, for example, comparisons
of sales by month.
Default An assumed value used by the DBMS unless specifically overridden.
Denormalization Adding redundant data to a table; fairly common in data warehouse
environments; also sometimes necessary in operational databases to improve the
performance of SELECT statements.
Disk drive A rotating storage device that is able to carry out one read or write
operation at a time.

Execution plan The output provided by the DBMS to describe an access path

being used.

Fact table Contains detailed transaction data, for example, sales or payment entries
such as sales figures, prices, or balances. This data is summarized or grouped with the
help of dimension data.
Fat table A table into which columns have been added from another table or which is
built by combining several tables together.
FETCH An SQL call used in cursor operations to request one row at a time.
Filter factor Specifies the selectivity of a predicate—what proportion of the table
rows satisfy the condition expressed by the predicate. It is dependent on the
distribution of the column values. When evaluating the adequacy of an index,
worst-case filter factors are more important than average filter factors.
Foreign key A column or column combination that points to the primary key in
another or same table. Foreign keys may have referential integrity constraints to
ensure data integrity.
Free space To cater for new rows being added to tables and indexes, a certain
proportion of each page may be left free when they are loaded or reorganized.