At the recent Strata conference, CitusDB showcased the latest release of its scalable analytics database built on top of PostgreSQL. The new release provides numerous new functionality and additional features.
According to CitusDB, CitusDB 2.0 brings together the performance of PostgreSQL and the scalability of Apache Hadoop, and enables real-time queries on data that's already in Hadoop. This new functionality is possible with CitusDB's distributed query planner, and PostgreSQL's foreign data wrappers.
Key new functionality provided in the new release includes network efficient LIMIT clauses, enabled by configuration, which allows limits on the amount of data being transferred over the network and provides better performance when using LIMIT clauses; smarter task assignment policy when shard count is low, which enables better cluster utilization when the shard count for a table is low compared to the number of workers; smarter foreign data wrapper integration for HDFS data; and support for sub-selects in the FROM clause which don't contain aggregate functions.
CitusDB 2.0 is based on PostgreSQL 9.2.1, and comes with all the features Postgres 9.2 brings, some of which include JSON data types; Range data types; and Index-only scans. A full list of Postgres 9.2 features can be found here. For more information, go here.