Transactional High Throughput Processing

This is an audit-less processing type, specifically designed to load large volumes of raw data as quickly as possible. It also supports DML statements.

Note:

If you convert a Transactional High Throughput data processing type Table into a Table with any of the other processing types (by modifying the Table's Process Type attribute), you are required to perform a full install of the Table when you install the Work Area next. This results in the re-creation of the Table and you lose all the existing data.

This processing type supports full and incremental data loading.

You can load large volumes of data by splitting the data loading job and yet be able to view all of the data together in a single snapshot.

This processing type has the following features:

• Displays All Data in a Single Snapshot. This data processing type makes all of the data available in a single snapshot even if the data is loaded through multiple jobs (when you use the Full data loading mode).

• Supports Blinding. This data processing type provides full support for data Blinding. You can mark the data loaded into the target Table of this processing type as Blinded or Dummy in the same way as with the other data processing types.

  Note:

  If the table is blinded and data has been loaded into both partitions then performing a load using full mode will delete the data from both partitions. You will lose all Blinded data if you choose the full mode for loading data, even if you run the job with the Dummy Blind Break setting. The system warns you if a new data-loading job is about to overwrite Blinded data giving you the option to cancel the job.

• Supports Full and Incremental Data Loading Modes. This processing type supports both incremental and full data loads. The default mode is full. In the full mode, the system truncates the existing Table and loads fresh data into it. In this mode, you will lose all the Blinded data in the Table, even if you run the job with the Dummy Blind Break setting.

  In the incremental mode, the new data is appended to the Table. The system hard-deletes data only if the Oracle LSH Program explicitly issues a Delete statement.

  Note:

  The default data loading mode is Full. If you do not want to lose all your existing data, change the data loading mode to Incremental.

• Supports Compression. If you create Oracle LSH Table instances marked with Transactional High Throughput data processing type in a tablespace that supports compression, you can compress these Tables.

• Supports Serialized Data Writing and Parallel Data Reading. Only one job can write data to a Table instance at a time. However, multiple jobs can read data from a Table instance, even when another writing job is running in parallel.

• Does Not Support Logical Rollback on Failure. In the event of a failure during data load, the system does not roll back the data that is already committed to the database. For example, if a job is inserting 5000 records into a Transactional High Throughput Table, and encounters an error at the 4000th record, the 3999 records already committed to the database are not rolled back.

  This is different from all the other data processing types, in that, when a job fails, the system removes all the data written to the Table and committed to the database as part of that job.

• Does Not Require Unique/Primary Key. You do not have to specify a unique/primary key for Table instances of this processing type. However if you specify these constraints, the Transactional High Throughput data processing type enforces them.