Graph databases speaking the same language after ISO gives GQL the nod

Standards body adoption could help ease portability between vendors

GQL, the query language for graph databases, has been recognized by the International Organization for Standardization (ISO), offering users more portability of queries and skills between graph database systems.

The first database query language to be certified by ISO since SQL in 1987, the new standard was developed by a group of database vendors, academics, and tech infrastructure companies.

Carl Olofson, research vice president at analyst IDC, said the industry had been waiting for "quite some time" for a standard graph query language to emerge.

"The formalization of GQL should remove a key barrier to graph database adoption," he said. "Although query languages are sometimes used for import and export, the main function of GQL is in traversing a graph with all its interconnections, finding patterns and locating specific nodes based on their relationships with other nodes. Other query languages, including SQL with graph enhancements, can't really support that completely."

The standard was edited by Stefan Plantikow, Neo4j product manager and standards engineer, and as such the graph database specialist is one of the leading companies in the collaboration. Neo4j also has its own query language called Cypher.

Jim Webber, Neo4j's chief scientist and a professor of computer science at Newcastle University, said those who understand the ubiquitous database query language SQL would see some familiar things in GQL, while other elements are very different.

"If you're familiar with SQL, then certainly things like the aggregate parts of GQL are going to look very familiar," he said. "But if you're familiar with things like Cypher or open Cypher, that's a much more straightforward segue into GQL. SQL deals with tables and sets, so its machinery is geared up for that. GQL code deals with graph so the machinery works on pattern matching and path expressions. Having said that, if you're someone that's competent at SQL, you're easily smart enough to be able to pick up [GQL] because graphs are a bit easier."

Webber argued that the emergence of the standard would give CIOs a sense of confidence in investing in the graph database model as they could more easily migrate their applications and queries from one vendor to another, so long as they have both adopted the standard.

Incidentally, GQL should not be confused with Google Query Language, used for Google's databases or GraphQL, a language for querying and managing APIs.

In a sign of harmony among software vendors, TigerGraph, Neo4j's main rival in the market, is also supporting the standard.

In a blog post, Mingxi Wu, TigerGraph SVP of engineering, writes: "As the graph database industry evolves, it becomes the third canonical database, alongside relational databases and key-value store databases.

"In the past decade, the burgeoning graph database industry has witnessed a plethora of vendors offering their own graph database products, each accompanied by their proprietary graph query language. Amidst this thriving ecosystem, GQL emerged to address the growing demand for a standardized graph query language. Its publication establishes a solid foundation and drives the prosperity of graph databases in the coming years, akin to what SQL did for relational databases."

Whether the graph database model is better for representing networks of relationships than relational systems is still up for debate. It did come out on top of The Register's Great Graph Debate last year, but only narrowly. A new standard query language might help, but it won't change the fundamentals. ®

More about

More about

More about


Send us news

Other stories you might like