Block Range Index

A Block Range Index or BRIN is a <a href="/facts/Database_index/TRqmk9It">database indexing</a> technique. They are intended to improve performance with extremely large tables.
BRIN indexes provide similar benefits to <a href="/facts/Horizontal_partitioning/2cKcwmXW">horizontal partitioning</a> or <a href="/facts/Sharding/YSE3YS0M">sharding</a> but without needing to explicitly declare partitions.
A BRIN is applicable to an index on a table that is large and where the index key value is easily sorted and evaluated with a MinMax function.
BRIN were originally proposed by Alvaro Herrera of 2ndQuadrant in 2013 as 'Minmax indexes'. Implementations thus far are tightly coupled to internal implementation and storage techniques for the database tables. This makes them efficient, but limits them to particular vendors. So far <a href="/facts/PostgreSQL/LHdXuHfl">PostgreSQL</a> is the only vendor to have announced a live product with this specific feature, in PostgreSQL 9.5. Other vendors have described some similar features, including <a href="/facts/Oracle_database/FHv5g732">Oracle</a>, <a href="/facts/Netezza/jVU2WM4f">Netezza</a> 'zone maps', <a href="/facts/Infobright/uPC3zmkB">Infobright</a> 'data packs', <a href="/facts/MonetDB/9DuGKGIN">MonetDB</a> and <a href="/facts/Apache_Hive/G3M79J6b">Apache Hive</a> with ORC/Parquet.

Block Range Index open-in-new

Block Range Index