Abstract: Techniques are described for performing partition pruning in situations where previous database servers could not. Specifically, techniques are provided for (1) performing partition pruning based on predicates that involve non-prefix subsets of composite partitioning keys, (2) performing partition pruning when a global index access path is used to execute a query, and (3) performing partition pruning when predicates on partitioning columns are OR-ed together. More generally, techniques are described for partition pruning based on a non-contiguous set of not-pruned partitions.

Abstract: The concept of portioning is expanded with a variety of techniques. In particular, one technique involves partitioning an object at multiple levels, where at least one of the levels uses list-based partitioning. Further, a partitioning technique is provided which involves storing a data item in a default partition when the item does not map to any one of the list-based partitions for which a key value group has been defined. Techniques are also provided for efficiently splitting a partition.

Abstract: The concept of portioning is expanded with a variety of techniques. In particular, one technique involves partitioning an object at multiple levels, where at least one of the levels uses list-based partitioning. Further, a partitioning technique is provided which involves storing a data item in a default partition when the item does not map to any one of the list-based partitions for which a key value group has been defined. Techniques are also provided for efficiently splitting a partition.

Abstract: Techniques for avoiding lock contention when processing data definition language (DDL) statements are provided. Some DDL statements modify metadata of a table and then require access to the table. After such a DDL statement modifies the metadata of a table and the updated metadata is made available (i.e., published) to other database statements, execution of the DDL statement is suspended. Data manipulation language (DML) statements acquire locks on the table. In one approach, the locks, of any DML statement, that are granted after the DDL statement is issued are timestamped. The DDL statement uses the timestamps to determine when to access the table. The timestamps are used to determine when the last DML statement (that was pending at the time the modified metadata was published) commits and releases its lock on the table.

Abstract: Techniques are described herein for partitioning a database table using a virtual column as the partitioning key. A virtual column is defined by the database to contain results of a defining expression. Partitioning can be performed using a variety of partitioning schemes, such as range partitioning and hash partitioning. The partitions are automatically updated to reflect changes to the values in the base columns upon which the virtual column depends. Partition pruning and parallel-wise partition joins can be performed on database tables which have been partitioned on virtual columns.

Abstract: Systems, methodologies, media, and other embodiments associated with supporting reference partitioned tables in a relational database are described. One example method includes identifying a partitioning strategy (e.g., range, list, hash) associated with a parent table that is related to a child table by a referential constraint. The example method may also include creating the child table to be a reference-partitioned table partitioned according to the partitioning strategy associated with the parent table. The example method may also include creating the child table to be equi-partitioned with respect to the parent table. The child table is not to replicate a partition key of the parent table but rather is to inherit a partitioning key associated with the parent table through a foreign key relationship.

Abstract: Techniques are provided for processing a CREATE INDEX statement that avoids one or more locks on the underlying table. The processing of such a statement may be divided into a build phase and a cooperation phase. During the build phase, a journal table is created and published. A snapshot of the table is obtained after the last DML transaction (that was pending at the time the journal table was published) commits. The index is built based on the state of the table as of the snapshot. Any changes to the table while the index is built are recorded in the journal table. During the cooperation phase, the index is partially published. A subsequent DML statement determines whether any changes in the journal table affect the same data as the subsequent DML statement. If so, then those changes are applied to the index followed by the subsequent DML statement.

Abstract: Techniques for avoiding lock contention when processing data definition language (DDL) statements are provided. Some DDL statements modify metadata of a table and then require access to the table. After such a DDL statement modifies the metadata of a table and the updated metadata is made available (i.e., published) to other database statements, execution of the DDL statement is suspended. Data manipulation language (DML) statements acquire locks on the table. In one approach, the locks, of any DML statement, that are granted after the DDL statement is issued are timestamped. The DDL statement uses the timestamps to determine when to access the table. The timestamps are used to determine when the last DML statement (that was pending at the time the modified metadata was published) commits and releases its lock on the table.

Abstract: Techniques for a data definition language (DDL) statement acquiring a lock on a table in wait mode in a deadlock free manner are provided. A DDL statement specifies metadata of a table that is to be modified, whereas a data manipulation language (DML) statement specifies a table to be modified. When a DDL statement is received, a proxy statement is initiated and executed instead of first acquiring an exclusive lock on metadata of a particular table. The proxy statement is similar to a DML statement in that the proxy statement acquires a lock on the particular table. After the proxy statement has been processed, an exclusive lock on the metadata is acquired.

Abstract: A method and apparatus for managing data in a partitioned data storage space is provided. Data referencing a new data item to be stored in the data storage space is received. It is determined that the data item does not pertain to any previously created partition. In response to the determination, a new partition is automatically created in the data storage space. The partition is mapped to a new set of values, such as to a next unpartitioned interval of dates or numbers.

Abstract: A method and apparatus for the on-demand allocation of segments and creation of metadata for previously-created data storage spaces and partitions are provided. A space is created in a database. As part of this creation process, no segment is allocated for the space. Rather, metadata describing the space sufficiently to allocate the segment in the future is created and maintained by the database. Data is received indicating a new item. Based on the metadata, it is determined that the new item pertains to the space. In response to the determination, a segment is allocated for the space. Additional metadata necessary for normal database operations in relation to the space and newly-allocated segment may also be created at this time. The new item can then be stored in the space.

Abstract: A method and apparatus for referencing a partition of a data storage space in a request to perform an operation with relation to the partition are disclosed. The techniques disclosed do not rely on knowing the partition's name, thus simplifying the process of requesting a database to perform an operation on a partition when that partition's name is unknown. Data such as a DDL statement is received. The data indicates a value in a FOR clause in lieu of the partition name. The data also indicates an operation to be performed with respect to a particular partition of the partitioned space. The particular partition is not named in the received data. Based on the value, it is determined that the particular partition is a first partition. The operation is performed with respect to the first partition.

Abstract: The concept of portioning is expanded with a variety of techniques. In particular, one technique involves partitioning an object at multiple levels, where at least one of the levels uses list-based partitioning. Further, a partitioning technique is provided which involves storing a data item in a default partition when the item does not map to any one of the list-based partitions for which a key value group has been defined. Techniques are also provided for efficiently splitting a partition.

Abstract: The concept of portioning is expanded with a variety of techniques. In particular, one technique involves partitioning an object at multiple levels, where at least one of the levels uses list-based partitioning. Further, a partitioning technique is provided which involves storing a data item in a default partition when the item does not map to any one of the list-based partitions for which a key value group has been defined. Techniques are also provided for efficiently splitting a partition.

Abstract: The concept of partitioning is expanded with a variety of techniques. In particular, one technique involves partitioning an object at multiple levels, where at least one of the levels uses list-based partitioning. Further, a partitioning technique is provided which involves storing a data item in a default partition when the item does not map to any one of the list-based partitions for which a key value group has been defined. Techniques are also provided for efficiently splitting a partition.

Abstract: Techniques are described for performing partition pruning in situations where previous database servers could not. Specifically, techniques are provided for (1) performing partition pruning based on predicates that involve non-prefix subsets of composite partitioning keys, (2) performing partition pruning when a global index access path is used to execute a query, and (3) performing partition pruning when predicates on partitioning columns are OR-ed together. More generally, techniques are described for partition pruning based on a non-contiguous set of not-pruned partitions.

Abstract: Each node in a data processing system contains at least one undo buffer and one least one redo buffer for insuring that any changes made to a section of a non-volatile storage medium, such as a disk, can be removed, if a transaction has not been committed, or can be recreated if the transaction has not been committed. The undo buffers each correspond to a different uncommitted transaction. The redo buffer contains the changes made to a copy of the section which is maintained in the memory.

Abstract: A data processing system maintains logs for system and media recovery. The logs contain state identifiers each uniquely identifying the state of a corresponding section of a storage medium, such as a disk. The state identifiers are assigned after changes have been made such that they can be determined from the information in the logs. One implementation involves assigning state identifiers in a known sequence, such as in a monotonically increasing sequence of integers beginning with zero.

Abstract: In a transaction processing system, snapshot records are stored in volatile memory along with volatile state memory records, and modifications to the volatile state memory records by the transactions are logged to an after-image log in non-volatile memory for recovery of the volatile state memory records. For recovery of the snapshot records in volatile memory, when any one of the records of volatile state memory are to be written from volatile memory to non-volatile state memory, the volatile snapshots records of the volatile state memory record are first written from volatile snapshot memory to non-volatile snapshot memory. This ordering of buffer pool flushing permits recovery of the volatile snapshot records from non-volatile state memory or from modifications in the after-image log. Moreover, recovery can be performed without writing modifications to the volatile snapshot records in an after-image log or flushing the snapshot records using an atomic operation.