Administration Guide

Populating a Table in a New Table Space

If you are populating a table in a new table space that does not yet contain data, do the following:

1. Execute db2split in analyze mode (to create a partitioning map), specifying a representative-sized data file for the largest table that will be in the table space. Analyze mode is established by setting the RunType parameter to ANALYZE in the configuration file.

2. Issue the REDISTRIBUTE NODEGROUP command, specifying the partitioning map created by db2split as the target partitioning map. This will put the partitioning map in the SYSPARTITIONMAPS catalog table.

3. Run db2split again, and have it partition the data according to the partitioning map that was just created. Partitioning is controlled by setting the RunType parameter to PARTITION in the configuration file.

4. Create the table in the table space with the CREATE TABLE statement.

5. Load the table on each database partition using the corresponding data files created during step 2. (The Load utility will check the partitioning map in the catalog table against the information in the data files that were split.)

6. Repeat steps three to five to split and load any other tables in the table space. When you do this, re-use the partitioning map that you created when you split the data of the first table.

[ DB2 List of Books | Search the DB2 Books ]