Partitioned Indexes

Partitioned Indexes

A partitioned index is an index that, like a partitioned table, has been decomposed into smaller and more manageable pieces. Global indexes are partitioned independently of the table on which they are created, whereas local indexes are automatically linked to the partitioning method for a table. Like partitioned tables, partitioned indexes improve manageability, availability, performance, and scalability.
The following graphic shows index partitioning options.

Local Partitioned Indexes

In a local partitioned index, the index is partitioned on the same columns, with the same number of partitions and the same partition bounds as its table. Each index partition is associated with exactly one partition of the underlying table, so that all keys in an index partition refer only to rows stored in a single table partition. In this way, the database automatically synchronizes index partitions with their associated table partitions, making each table-index pair independent.
Local partitioned indexes are common in data warehousing environments. Local indexes offer the following advantages:
■Availability is increased because actions that make data invalid or unavailable in a partition affect this partition only.
■Partition maintenance is simplified. When moving a table partition, or when data ages out of a partition, only the associated local index partition must be rebuilt or maintained. In a global index, all index partitions must be rebuilt or maintained.
■If point-in-time recovery of a partition occurs, then the indexes can be recovered to the recovery time (see "Data File Recovery" on page 18-14). The entire index does not need to be rebuilt.
Example 4–4 shows the creation statement for the partitioned hash_sales table, using the prod_id column as partition key. Example 4–5 creates a local partitioned index on the time_id column of the hash_sales table.

Example 4–5 Local Partitioned Index
CREATE INDEX hash_sales_idx ON hash_sales(time_id) LOCAL;

In Figure 4–4, the hash_products table has two partitions, so hash_sales_idx has two partitions. Each index partition is associated with a different table partition. Index partition SYS_P38 indexes rows in table partition SYS_P33, whereas index partition SYS_P39 indexes rows in table partition SYS_P34.

You cannot explicitly add a partition to a local index. Instead, new partitions are added to local indexes only when you add a partition to the underlying table. Likewise, you cannot explicitly drop a partition from a local index. Instead, local index partitions are dropped only when you drop a partition from the underlying table.
Like other indexes, you can create a bitmap index on partitioned tables. The only restriction is that bitmap indexes must be local to the partitioned table—they cannot be global indexes. Global bitmap indexes are supported only on nonpartitioned tables.

Global Partitioned Indexes

A global partitioned index is a B-tree index that is partitioned independently of the underlying table on which it is created. A single index partition can point to any or all table partitions, whereas in a locally partitioned index, a one-to-one parity exists between index partitions and table partitions.
In general, global indexes are useful for OLTP applications, where rapid access, data integrity, and availability are important. In an OLTP system, a table may be partitioned by one key, for example, the employees.department_id column, but an application may need to access the data with many different keys, for example, by employee_id or job_id. Global indexes can be useful in this scenario.
You can partition a global index by range or by hash. If partitioned by range, then the database partitions the global index on the ranges of values from the table columns you specify in the column list. If partitioned by hash, then the database assigns rows to the partitions using a hash function on values in the partitioning key columns.
As an illustration, suppose that you create a global partitioned index on the time_range_sales table from Example 4–2. In this table, rows for sales from 1998 are stored in one partition, rows for sales from 1999 are in another, and so on. Example 4–6 creates a global index partitioned by range on the channel_id column.

Example 4–2 Range-Partitioned Table
CREATE TABLE time_range_sales
( prod_id NUMBER(6)
, cust_id NUMBER
, time_id DATE
, channel_id CHAR(1)
, promo_id NUMBER(6)
, quantity_sold NUMBER(3)
, amount_sold NUMBER(10,2)
)
PARTITION BY RANGE (time_id)
(PARTITION SALES_1998 VALUES LESS THAN (TO_DATE('01-JAN-1999','DD-MON-YYYY')),
PARTITION SALES_1999 VALUES LESS THAN (TO_DATE('01-JAN-2000','DD-MON-YYYY')),
PARTITION SALES_2000 VALUES LESS THAN (TO_DATE('01-JAN-2001','DD-MON-YYYY')),
PARTITION SALES_2001 VALUES LESS THAN (MAXVALUE)
);

Example 4–6 Global Partitioned Index
CREATE INDEX time_channel_sales_idx ON time_range_sales (channel_id)
GLOBAL PARTITION BY RANGE (channel_id)
(PARTITION p1 VALUES LESS THAN (3),
PARTITION p2 VALUES LESS THAN (4),
PARTITION p3 VALUES LESS THAN (MAXVALUE));

As shown in Figure 4–5, a global index partition can contain entries that point to multiple table partitions. Index partition p1 points to the rows with a channel_id of 2, index partition p2 points to the rows with a channel_id of 3, and index partition p3 points to the rows with a channel_id of 4 or 9.