We are very excited to announce the release of AgensGraph v1.1. It includes both new features and many useful improvements. Since we released AgensGraph v1.0, our team has been hard at work researching and building the best AgensGraph solution for database over the past two months. With improved features and upgraded performance, AgensGraph makes it easier for you to manage, analyze and store your data.
Let’s take a look at the new features and improvements:
The MERGE of OpenCypher (http://opencypher.org) specification is a very convenient way to ensure the existence of specified patterns in a graph. If there is no such pattern, AgensGraph creates the specified pattern automatically. Users can also specify the additional behaviors to be executed according to the pattern existence using ON MATCH and ON CREATE optional clauses. Please refer to the details in the AgensGraph documentation (https://bitnine.net/support/quick-start-guide-html/#merge).
Performance Improvements of Variable Length Edge Pattern Queries
AgensGraph v1.1 improves the performance for variable length edge pattern queries by more than two times. In OpenCypher, users can specify a path pattern which has variable length edges by using
* notation in edge patterns, e.g.
(a)-[*2]-(b). This feature is a very convenient and concise way to specify degrees of relationships between entities such as finding friends of a friend. In a relational database, the searches can usually be described as recursive common table expressions. However, this makes queries less readable and hard to maintain. AgensGraph has supported this variable length edge pattern since v1.0 and in v1.1 The performance has improved dramatically. This is one of the reasons why users choose graph database over relational database.
PostgreSQL 9.6.2 Update
The PostgreSQL Global Community announced the release of PostgreSQL 9.6.2 in February 2017. The recent release of PostgreSQL has been integrated into AgensGraph v1.1. The details of PostgreSQL 9.6.2 can be found in https://www.postgresql.org/docs/current/static/release-9-6-2.html. Bitnine developer, Junseok Yang, also contributes explanation about the patch of the array on PostgreSQL 9.6.2 in the previous link. AgensGraph team will continue to maintain PostgreSQL code base and contribute to the global community. By updating the code base, AgensGraph becomes more reliable. This is an important benefit of AgensGraph.
In AgensGraph v1.1, two external modules have been added: Hadoop FDW and HyperLogLog datatype.
Hadoop FDW is from OpenSCG (http://hadoopfdw.org/) and it is a foreign data wrapper which connects HIVE systems to access HDFS files in AgensGraph. Using Hadoop FDW, the data residing in Hadoop can be easily imported into AgensGraph.
HyperLogLog (https://github.com/aggregateknowledge/postgresql-hll) is an external module by Neustar, which can be used to estimate the distinct value count using a single scan and small amount of memory without keeping all distinct values in memory. This module can be used to help AgensGraph’s query optimizer by providing more exact statistics for the distinct value estimation.
AgensGraph v1.1 is a minor release but has major features and improvements. The AgensGraph Team will continue to cooperate and contribute to PostgreSQL’s global community in an effort to improve the core engine. In addition, the support for Big Data ecosystems will be continuously improved. Please refer to https://github.com/bitnine-oss/agensgraph/wiki for more detailed information.
BITNINE GLOBAL INC., THE COMPANY SPECIALIZING IN GRAPH DATABASE
비트나인, 그래프 데이터베이스 전문 기업