Defining partitioned databases

You define partitioned databases in the DBD by:

1. Replacing the first DATASET statement with a BMC AMI Partition Facility for IMS PART statement—using one PART statement for each partition. If you will not be using data set groups, remove any other DATASET statements. You are not required to change any other DBD statements to partition a database.
2. Adding PARTDBD and/or DATASET statements if using partition selection routine or data set groups.
3. Running DBDGEN.
4. Running ACBGEN.
5. If you register the database with DBRC, you must also add INIT.DBDS definitions and modify the JCL DD statements and/or dynamic allocation members for the additional data sets.

BMC AMI Partition Facility for IMS allows you to define partitioned databases with only one partition. Why would you want to define a single partition database? One reason might be to perform a two-step migration — converting the database to a single partition database, and then adding more partitions as the database grows.

To create a single-partition database you use a single PART statement in the DBD instead of a single DATASET statement. All restrictions for multiple-partition databases apply to single-partition databases.

With single partition databases, primary or secondary indexes are not required to indicate in which partition the target exists, because the RBA is always in the first partition, and there is no partition HIGHKEY with a single partition HIDAM database.

This section explains how to define partitioned databases. Valid database organizations are HDAM, HIDAM, HISAM, SHISAM, and INDEX. This section contains the following topics: