thumb|Partitioning options on a table in MySQL in the environment of the [[Adminer tool.]]
A partition is a division of a logical database or its constituent elements into distinct independent parts. Database partitioning refers to intentionally breaking a large database into smaller ones for scalability purposes, distinct from network partitions which are a type of network fault between nodes. In a partitioned database, each piece of data belongs to exactly one partition, effectively making each partition a small database of its own.
use varying terminology for partitioning:
- Shard in MongoDB, Elasticsearch, and SolrCloud
- Region in HBase
- Tablet in Bigtable
- vnode in Cassandra and Riak
- vBucket in Couchbase
Generally, this practice is known as normalization. However, vertical partitioning extends further, and partitions columns even when already normalized. This type of partitioning is also called "row splitting", since rows get split by their columns, and might be performed explicitly or implicitly. Distinct physical machines might be used to realize vertical partitioning: storing infrequently used or very wide columns, taking up a significant amount of memory, on a different machine, for example, is a method of vertical partitioning. A common form of vertical partitioning is to split static data from dynamic data, since the former is faster to access than the latter, particularly for a table where the dynamic data is not used as often as the static. Creating a view across the two newly created tables restores the original table with a performance penalty, but accessing the static data alone will show higher performance. A columnar database can be regarded as a database that has been vertically partitioned until each column is stored in its own table.
See also
- Block Range Index
- CAP theorem
- Data striping in RAIDs