Partitioning DL/1 databases

DL/1 supports multiple data sets with data set groups, but requires all segments of the same type to be in the same data\ set.

Data set groups allow a dependent segment to be stored in a separate data set from the root. This can be useful for large, infrequently accessed segment types. An extra I/O is needed to read the entire database record since the dependent segment is in a separate data set.

Data set groups can are useful for databases greater than 4 GB when one or two segments comprise a significant percentage of the database, and the segments are not frequently accessed or updated. However, data set groups are not always a permanent solution to the problems of large databases.

Creating Data Set Groups

Partitioned databases implement multiple database data sets similar to DEDB Areas.

All segments in the same physical database record are in the same partitioned data set. With both HDAM and HIDAM, the partitioning is done by the root key. All physical dependents are stored in the same data set as the root. A database that is partitioned uses at least one, primary data set group. A partitioned database can have more than one data set group, but no more than 10 data set groups. The number of data set groups multiplied by the number of partitions must not exceed 127. For example, if you designate 31 partitions, you can have no more than four data set groups.

Creating Partitions