Data Snapshots

Snapshots allow Oracle LSH to view the data in one or more Table instances in the state it was in at the completion of any job that modified data in the Table instance(s).

The system uses the master job's refresh timestamp for the creation timestamp (and, for deleted records, end timestamp) of all records processed in a master job.

The system uses the master job's refresh timestamp for the creation timestamp (and, for deleted records, end timestamp) of all records processed in a master job, even if the job takes place over a number of hours and includes incremental commits. A master job is any job submitted explicitly for execution. Some master jobs include subjobs; for example, Workflows, Report Sets and any executable submitted using backchaining.

A snapshot comprises all the records in a Table instance that are current at a given point in time. For reload and audited transaction Table instances, a snapshot is the set of records whose end timestamp is greater than, and whose creation timestamp is less than or equal to, a given refresh timestamp. For audited staging Table instances, it is the set of records whose creation timestamp equals the refresh timestamp.

Using snapshots has the following benefits:

• When Programs subsequently access the resulting data, the system bases a stable view of the data on the most recent refresh timestamp, providing a consistent view of the data even if the table is being updated at the time the Program reading the data is running.

• It is possible to recreate data as it was at an earlier point in time.

• Access to incompletely loaded data is prevented and rollbacks of incomplete loads are supported even if there have been incremental commits.

You can label snapshots in two ways:

• When you run a job you can specify a label to be applied to the source and/or target Table instances (see instructions for "Generating Reports and Running Other Jobs" in the Oracle Life Sciences Data Hub User's Guide).

• In a Work Area, you can apply a snapshot label to a data timestamp in one or more Table instances in the Work Area; see Adding, Removing, or Moving a Snapshot Label.