To identify tables for OpenEdge table partitioning, you must make a number of decisions. Since your application developers are typically knowledgeable about the logical design of a database and how it is used by their applications, you should work closely with them in making these decisions. Specifically, you need to answer the following questions:

  • Which of the very large tables in a database must be partitioned?
  • Which columns of the tables should be used as partition keys?
  • What type of partition should be used for each partitioned table?
  • If you use range partitioning or subpartitioning that includes range, where should the partition boundaries be drawn?
    • In other words, how many partitions should be allocated to each partitioned table?
  • How should the partitions be stored?
    • Should each partition have its own storage area?
    • Should two or more partitions be in the same storage area?
  • What kinds of indexes should be used for each partitioned table?
    • Which indexes should be local?
    • Which indexes should be global?

After you answer all these questions, you are ready to allocate space in the database for the tables that you are going to partition.