This site is about a previous instance of this event. For the current event, please see the updated site.

Schedule - PGConf.EU 2022

Efficient Graph Analysis with SQL/PGQ

Date: 2022-10-26
Time: 09:50–10:40
Room: Berlin 1
Level: Beginner

The upcoming SQL:2023 standard will bring as main new addition the ability to perform queries on graph data: Property Graph Query (PGQ) is a new sub-language of SQL that adds MATCH functionality for pattern matching and path-finding in "property graphs" (similar to neo4jā€™s Cypher query language). The CREATE PROPERTY GRAPH statement creates a property graph view on relational data, where the vertexes and edges (and their properties) are both represented by normal SQL tables. SQL/PGQ is part of the larger GQL language, also under development by the SQL working group of ISO, in liaison with a non-profit organization called LDBC (, which I founded and still chair, that facilitates standards development for graph data management. Through LDBC it is possible to get access to the working specs of SQL/PGQ. The next step is implementing this new functionality in database systems. My database group at CWI is known for its pioneering work on techniques like vectorized execution, and skippable columnar storage with lightweight compression, adopted by most analytical database systems and data formats like Parquet and Arrow. CWI produced systems like MonetDB, VectorWise, and now the highly popular DuckDB, for embeddable analytics; in the process also founding numerous spin-off companies, and creating many connections with industry (e.g. Snowflake and Databricks). In the keynote I will also reflect on CWI's ongoing SQL/PGQ implementation in DuckDB, as well as on the (im)possibility for efficient graph analytics in PostgreSQL.


The following slides have been made available for this session:


Peter Boncz