Besides overall size, one should look at the data access patterns. It is helpful if the table is partitioned on how it is accessed. For instance, if I have a table for INVOICES and the invoices are typically searched by DATE, then I would partition on the invoice date column. Each partition might hold a year's worth, or a quarter's worth of data. By partitioning with the data access in mind, one can enjoy the benefits of "partition pruning." This is eliminating those partitions that will not participate in the query. So when I query for all invoices generated last year, only the partition(s) that hold that data will be used. Partitions holding previous year's data will be ignored.
Dig Deeper on Oracle database design and architecture
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.