Knowledge Tree
Property Graph Model
Also known as: PGM, Property Graph Models, Property Graph Database Model
Facts (46)
Sources
Construction of Knowledge Graphs: State and Challenges - arXiv arxiv.org 42 facts
claimProperty Graph Models serve as the base data model for graph query languages including G-Core, Gremlin, PGQL, Cypher, SQL/PGQ, and GQL.
claimRDF has undergone extensive standardization over the last 25 years, whereas Property Graph Models (PGM) have become increasingly popular for advanced database and network applications like graph traversal and network analysis.
claimCommon implementations of Property Graph Models specify vertices and edges using unique identifiers.
claimData integrity in Property Graph Model databases is generally limited to syntax or basic value constraints, similar to RDF stores.
referenceO. Hartig proposed a method for the reconciliation of RDF* and property graphs in a 2014 CoRR paper.
claimThe communities surrounding Property Graph Models (PGM) and RDF use many similar but differently named terms, leading to terminology variations.
claimProperty Graph Models intentionally lack a predefined schema to facilitate the flexible incorporation of heterogeneous entities and relations, though schema graphs can be inferred from type information.
claimThe choice between using RDF, Property Graph Models (PGM), or a custom data model depends on the targeted application or use case of the final knowledge graph.
referenceG. Abuoda et al. conducted a preliminary analysis of approaches for transforming RDF-star to property graphs in a 2022 arXiv preprint.
claimThe authors of 'Construction of Knowledge Graphs: State and Challenges' expand the scope of knowledge graph construction research to include non-RDF-based models like the Property Graph Model, the integration of structured and unstructured data, and incremental maintenance.
claimThe Property Graph Model (PGM), also known as Labeled Property Graph (LPG), supports the definition of graph structures with heterogeneous nodes and directed edges to represent entities and their relationships.
claimThe authors of the survey 'Construction of Knowledge Graphs: State and Challenges' claim their work is more comprehensive than existing surveys because it covers both RDF and property graph data models, as well as incremental knowledge graph construction and incremental entity resolution.
claimThe Amazon Neptune database service allows users to operate Property Graph Models (PGM) and RDF interchangeably.
claimThe capabilities of Property Graph Models depend heavily on their specific implementation because there is no global de facto standard for the model.
claimThe graph query language GQL will support a feature for the Property Graph Model (PGM) similar to features found in other graph query languages.
claimA benchmark for knowledge graph construction should ideally involve the initial construction and incremental update of domain-specific or cross-domain knowledge graphs from diverse data sources, using predefined ontologies and data models like RDF or property graphs to facilitate evaluation.
measurementThe CovidGraph knowledge graph, established in 2020, contains 36 million entities and 59 million facts, utilizing PGM format.
claimThe most common graph models used for knowledge graphs are the Resource Description Framework and the Property Graph Model.
referenceThe paper 'Schema inference for property graphs' by H. Lbath, A. Bonifati, and R. Harmer was presented at the 24th International Conference on Extending Database Technology (EDBT) in 2021.
referenceThe Property Graph Exchange Format is a format for exchanging property graph data, described in an ArXiv preprint.
claimRDF-Star improves the formal meta-expressiveness of RDF, but specific cases still require support constructs to be presentable, unlike in Property Graph Models (PGM).
perspectiveLassila et al. conclude that both RDF and Property Graph Models (PGM) are qualified to meet their respective challenges but neither is perfect for every use case, recommending increased interoperability between both models to reuse existing techniques.
claimProperty Graph Models allow for the maintenance of embedded metadata for entities and relationships through the use of dedicated properties, such as those for provenance or time annotations.
claimProperty Graph Models are supported by graph database systems including Neo4j, JanusGraph, and TigerGraph, as well as processing frameworks like Oracle Labs PGX and Gradoop.
claimExtensions to the Property Graph Model exist to support temporally evolving graph data and graph streams, often providing advanced analysis capabilities for graph mining.
referenceCypher is an evolving query language for property graphs, presented at the 2018 International Conference on Management of Data (SIGMOD Conference 2018).
referenceA. Bonifati et al. defined 'PG-Schema', a schema framework for property graphs, in a 2022 paper published in the Proceedings of the ACM on Management of Data.
claimEfforts to create a standardized Property Graph Model serialization format include the JSON-based Property Graph Exchange Format (PGEF), YARS-PG, and the Graph Definition Language (GDL).
referenceR. Angles, H. Thakkar, and D. Tomaszuk analyzed the status and issues regarding interoperability between RDF and Property Graphs in 2019.
claimProperty Graph Models are more closely related to graph models in graph theory than RDF, which contributes to their understandability.
referenceR. Angles et al. introduced 'PG-Keys', a system for defining keys for property graphs, in a 2021 paper presented at the International Conference on Management of Data.
claimThe smallest unit of information is defined as a statement or fact; for RDF this describes a triple, while for Property Graph Models (PGM) this can be assigning a property-value, adding a type label to an entity, or adding a relation between two nodes.
claimGraphQL provides a unified approach to query both RDF and Property Graph Models (PGM), though it offers fewer features compared to query languages dedicated to these specific graph formats.
referenceThe paper 'The Property Graph Database Model' by R. Angles was presented at the AMW conference in 2018.
claimProperty Graph Models lack built-in support for ontologies, such as providing 'is-a' relations between entity categories.
claimProperty Graph Models allow nodes and edges to have multiple type-labels and key-value property pairs.
claimProperty Graph Models (PGM) allow for two relations with the same name to be addressed independently, with each relation having its own distinct properties.
claimKnowledge graph construction pipelines must transform input data into a final format, such as RDF or a property graph, often requiring multiple format conversions between pipeline steps.
referenceSzeremeta, K. Litman, and D. Cisterna proposed a serialization method for property graphs in their 2019 paper 'Serialization for Property Graphs', presented at the 15th International Conference on Beyond Databases, Architectures and Structures (BDAS 2019).
claimThere has been relatively little research investigating the transformation of structured data into property graphs compared to RDF-based approaches.
referenceHartig et al. and Abuoda et al. discuss transformation strategies between RDF and Property Graph Models (PGM) to lower usage boundaries.
referenceAngles et al. proposed four natural key types for property graph constraints: identifier, exclusive mandatory, exclusive singleton, and exclusive.
Context Graph vs Knowledge Graph: Key Differences for AI - Atlan atlan.com Jan 27, 2026 2 facts
claimKnowledge graphs are built on RDF triple stores or property graphs like Neo4j, whereas context graphs are built on graph databases extended for operational and AI context.
referenceKnowledge graphs are queried using SPARQL for triple stores or Cypher for property graphs, whereas context graphs utilize graph queries combined with operational filters to find assets based on quality, certification, and modification history.
Bridging the Gap Between LLMs and Evolving Medical Knowledge arxiv.org Jun 29, 2025 1 fact
referenceThe Neo4j property graph model consists of nodes (entities representing data points), relationships (directed, named connections between nodes), and properties (key-value pairs associated with nodes and relationships).
Efficient Knowledge Graph Construction and Retrieval from ... - arXiv arxiv.org Aug 7, 2025 1 fact
procedureThe GraphProducer module accepts generic triples generated by the model and transforms them into a property graph format compatible with the target graph database, with future plans to support RDF triple conversion.