Current location - Loan Platform Complete Network - Big data management - Must oracle build partitions for large data tables?
Must oracle build partitions for large data tables?
The recommendation given by oracle is given according to the size of the table, and I remember that the recommendation for 10g is 2G, which means that if the size of the table is larger than 2G, then it is better to create partitions. \x0d\ Why do we need to create partitions? Mainly for the convenience of querying because if a table has millions of records every day, then even if an index is created, then you won't be able to check it very fast if you want to. \x0d\ If you create partitions, each time you query a partition, which is a fraction of this table, or even a few tenths of it, then no matter how you calculate it is less consuming than if you query the whole table. And you can create indexes on partitions as well. \x0d\ Partitioning the establishment of the best a principle, is the query, this is a specific problem to analyze, some partitioning, although the number of entries is good, however, each query to cross 2-3 partitions, such partitioning actually personally believe that more of a failure. It is best to limit the most commonly used queries in a partition, and the number of entries in the partition can not be too small (this is not a specific requirement, the general oracle is in accordance with the size of the proposed, I forget how many, or that, depending on your needs)\x0d\but in accordance with your question, the big data table (the kind of data volume is very large), it is best to establish a partition so that the check up is convenient, save resources and time.