A composite partition is a physical partition where more than one logically defined partition resides in the same physical storage allocation. Prior to the migration of data in an existing table into partitions, the partition containing the table in its pre-partitioned state is a composite partition.
Global index
A global index in an index that contains key entries for data across all partitions of the partitioned table. Global indexes are used when a sort order other than the partitioned aligned sort order is required. For example, an index for customer name when the table is partitioned by region.
Horizontal data/table partitioning
Horizontal data/table partitioning is the partitioning of a table based on values of specified columns. Each row of a partition contains all the columns of the table but only contains those rows where the column value meets a particular partition policy established by the DBA.
Index of indexes
The index of indexes is an index b-tree that is special for table partitioning. The index of indexes is built and scanned like a regular index but the entries in the leaf nodes point to the root block of a local index, making it an “index of indexes” rather than an index of record pointers. The index of indexes is used to facilitate partition pruning for the languages without the need for special knowledge of the partitions being scanned. The index of indexes is also used to provide partition information for table scans of partitioned tables.
List partitioning
List partitioning partitions data from your table based on a defined set of values. For example, partitioning a table by country or state. If you have list partition based on the value of a country field, it requires a partitioning policy that lists the partition information for data storage of each valid country field value such as USA, Canada, Mexico, etc.
Local index (and local –prefixed (or aligned) indexes)
A local index is an index that contains key entries only for one particular table partition.
Multi-column partitioning
Multi-column partitioning is equivalent to sub-partitioning.
Partition aligned indexes (or partition prefixed)
Partition aligned indexes are indexes that include as the leading component(s) the column(s) associated with and in order of the table’s partition policy (the partition key).
Partition id
A partition id is database-wide unique identifier associated with each partition policy detail record in the database which is used to locate the physical data associated with the partition policy.
Partition Key
The partition key is the ordered column or columns that uniquely identify the table partition. The partition key is described in the partition policy.
Partition name
A partition name is a (database wide) unique name given to a table partition policy. A partition name can be either system generated implicitly or specified by the user explicitly. A valid partition name follows the same rules as a valid table name.
Partition pruning
Partition pruning is the process of analyzing a request for data such that only the partitions containing data associated with the request are examined, For instance, for each sales where quarter GT EQ 3 and year EQ 2011: only looks at 3 partition out of the all the partitions associated with the particular table. Partition pruning is performed by the storage engine when no specific partition is requested by the language engines.
Physical data partition
The physical data storage partition is the “container” associated with a particular table partition policy.
Range partitioning
Range partitioning is the partitioning of a table based on a value range associated with a particular column. The column can be any indexable data type that you can divide with a greater than or less than (<=) evaluation. Simple examples include partition definitions based on a range of values for the given data type such as a historical (date) range or a salary (numeric) range.
Sub-partitioning
Sub-partitioning is partitioning of more than one column of the same table (multi-column partitioning). Some examples are range-list partitioning or list-list partitioning. It is the combination of the values of each of the participating columns which determine the table partition to use for the data.
Table Partition Policy
A table partition policy is a policy record stored as database meta-schema that defines how a particular table is partitioned. A partition policy includes the partition name, the partition type, the column or columns involved and various attributes associated with the partition.
Table Partition Policy Detail
The partition policy detail is a policy record stored as database meta-schema that defines the list of values or ranges for each column defined in the table partition policy.
Table Partitioning
Table partitioning is the act of splitting data associated with a table definition into separately managed pieces, based on the value or range of values for a specific column or columns.
