|Couchbase Launches JSON Analytics|
|Written by Kay Ewbank|
|Friday, 28 September 2018|
There's a new analytics service for Couchbase that comes with SQL++, a query language framework that can be used to write queries on document-oriented data and the JSON data format.
The new analytics service is JSON-native, and designed to avoid the need for an ETL transformation when bringing data into the NoSQL store.
CouchDB was originally developed by Damien Katz, and was then adopted as a successful Apache project. The developers then moved on to create a successor to CouchDB, Couchbase Server. This is a hybrid database combining elements of both document and key-value databases. The key-value element comes from the memcached-compatible protocol, while documents are handled by storing the data in JSON format with support for secondary indexes.
According to the developers, the JSON-native analytics engine will let users perform parallel ad-hoc analytics on data as soon as it's been loaded into the NoSQL store. Under the covers, the new Analytics Service makes use of features from Apache AsterixDB, as well as academic research on SQL++. Extract, Transform and Load (ETL) can be a major operation when trying to make data available for use, and the act of taking complex but loosely structured data from JSON documents into the more rigid requirements of traditional analysis services can lose a lot of information.
Couchbase's approach attempts to keep the JSON data structure are based on research from the University of California on AsterixDB, and also on a modified take on Hybrid Transactional Analytical Processing (HTAP), a technique proposed by Gartner for data analysis carried out within production systems rather than the more usual move to a separate system for analysis.
The new Couchbase service is designed to let business users run ad-hoc analytical queries by using a massively parallel processing (MPP) architecture to avoid having an unacceptable impact on the performance of the main database. It supports ad-hoc querying of data, and the developers say data is available for analytical processing in milliseconds. The query language, which the company is calling N1QL for Analytics, is an implementation of the SQL++ language for querying schema-less semi-structured JSON data. The development of the SQL++ framework is based on UC San Diego’s NSF-funded FORWARD project with subsequent contributions and funding from Couchbase and Informatica. The developers say that as the language is derived from SQL++, N1QL for Analytics is easy to learn for developers who are familiar with the syntax of SQL.
or email your comment to: firstname.lastname@example.org
|Last Updated ( Friday, 28 September 2018 )|