This preface contains:
Changes in Oracle Database 12c Release 2 (188.8.131.52)
Oracle Database In-Memory Guide for Oracle Database 12c Release 2 (184.108.40.206) has the following changes.
The following major features are new in this release:
In-Memory Column Store (IM column store) dynamic resizing
You can now dynamically increase the size of the In-Memory Area without reopening the database.
In-Memory Expressions (IM expressions)
Oracle Database automatically identifies frequently used (“hot”) expressions that are candidates for population in the IM column store. A candidate expression might be
(monthly_sales*12)/52. IM expressions can greatly improve the performance of analytic queries that use computationally intensive expressions and access large data sets.
In-Memory virtual columns (IM virtual columns)
IM virtual columns enable the IM column store to materialize some or all virtual columns in a table.
IM FastStart optimizes the population of database objects in the IM column store by storing IMCUs directly on disk.
Object-level support for services
For an individual object, the
INMEMORY ... DISTRIBUTEclause has a
FOR SERVICEsubclause that limits population to the database instance where this service can run. For example, you can configure an
INMEMORYobject to be populated in the IM column store on instance 1 only, or on instance 2 only, or in both instances.
IM column store on a standby database
You can enable an IM column store in an Oracle Active Data Guard standby database. You can populate a completely different set of data in the in-memory column store on the primary and standby databases, effectively doubling the size of the in-memory column store that is available to the application.
ADO support for the IM column store
You can use Automatic Data Optimization (ADO) policies to evict objects such as tables, partitions, or subpartitions from the IM column store based on Heat Map statistics. Successful policy completion results in setting
NO INMEMORYfor the specified object.
A join group is a user-created object that lists two columns that can be meaningfully joined. In certain queries, join groups enable the database to eliminate the performance overhead of decompressing and hashing column values. Join groups require an IM column store.