5.1.7 Data loading with the synchronization point specification

When loading data, usually a transaction is settled after all data has been stored. If the utility terminates abnormally during the processing, the transaction rolls back to the start point, in which case data loading must be re-executed from the beginning.

The data loading method with the synchronization point specification enables a transaction to be settled each time a specified number of data items has been stored. This method is suitable for loading a large amount of data because it reduces the rollback time, as well as the data loading time, in the event of an abnormal termination of the utility.

Data loading with synchronization point specification is not applicable to index creation (batch index creation mode) or LOB column data loading. In such cases, the transaction is settled after all processing is completed.

To execute data loading with the synchronization point specification, specify the option statement (job operand). The figure below provides an overview of data loading with the synchronization point specification.

Figure 5-5 Overview of data loading with the synchronization point specification

[Figure]

Explanation
At the initial execution, the transaction rolls back after the occurrence of an error. This rollback is to the point of two million data items because the transaction has settled at that point.
The re-execution skips to the point of two million data items and starts storing the remaining data.