Distributed files with DB2^® Multisystem

A distributed file is a database file that is spread across multiple iSeries™ servers.

Distributed files can be updated, and they can be accessed through such methods as SQL, query tools, and the Display Physical File Member (DSPPFM) command. For database operations, distributed files are treated just like local files, for the most part. For information about CL command changes for distributed files, see How CL commands work with distributed files topic.

To create a database file as a distributed file, you can use either the Create Physical File (CRTPF) command or the SQL CREATE TABLE statement. Both methods are discussed in more detail in this topic. The CRTPF command has two parameters, node group (NODGRP) and partitioning key (PTNKEY), that create the file as a distributed file. You can see the distributed database file as an object (*FILE with the PF attribute) that has the name and library that you specified when you ran the CRTPF command.

If you want to change an existing nondistributed database physical file into a distributed file, you can do this by using the Change Physical File (CHGPF) command. On the CHGPF command, the NODGRP and PTNKEY parameters allow you to make the change to a distributed file. Through the CHGPF command, you can also change the data partitioning attributes for an existing distributed database file by specifying values for the NODGRP and PTNKEY parameters. Specifying values for these parameters causes the data to be redistributed according to the partitioning table in the node group.

• Create Physical File (CRTPF) command and SQL CREATE TABLE statement
  This topic provides information about creating a distributed physical file by using the Create Physical File (CRTPF) command or the SQL CREATE TABLE statement.
• System activities after the distributed file is created
  As soon as the file is created, the system ensures that the data is partitioned and that the files remain at concurrent levels.
• Partitioning with DB2 Multisystem
  Partitioning is the process of distributing a file across the nodes in a node group.
• Customization of data distribution with DB2 Multisystem
  Because the system is responsible for placing the data, you do not need to know where the records actually reside. However, if you want to guarantee that certain records are always stored on a particular system, you can use the Change Node Group Attributes (CHGNODGRPA) command to specify where those records reside.