What does partitioning mean in regards to a database?

Partitioning a database means taking various parts of the data stored in the database and separating them into various partitions, or pieces. This is often done to accommodate load balancing, or to help provide smaller database sets that can be worked on by independent server systems.

In a partitioned database, some partitioned sets may be shared across various nodes or servers. Others will be housed independently in one node.

Different types of partitioning included range partitioning, list partitioning and hash partitioning. In range partitioning, engineers partition a single database table by splitting it up into multiple key ranges. List partitioning involves taking a top-level identifier and grouping data sets into a partitioned database node. Hash partitioning uses a hash function to sort out the partitioning of data sets.

Engineers also talk about “shards” in horizontal partitioning. A database shard is held in its own server to, again, accommodate load balancing or load sharing. Database partitioning is a growing phenomenon as relational databases and other tools handle increasing volumes of data in the age of big data and analytics.