Compatibility and capabilities

This page distinguishes three different claims:

client compatibility comes from the database vendor;
implemented means Nodal has a provider implementation and automated compiler or contract tests;
live verified means the repository exercises the feature against the exact database baseline listed below.

Nodal does not infer that a feature exists merely because a driver can connect to a server version.

Version policy

Component	Nodal baseline	Vendor compatibility	Current verification
.NET	.NET 10	Nodal packages currently target `net10.0`	Build, unit tests, package validation
Neo4j driver	`Neo4j.Driver` 6.3.0	Driver 6.x supports Neo4j 4.4.x, 5.x, 2025.x, and 2026.x	Package and compiler tests
Neo4j database	Neo4j 5.26 Community	Wider connectivity follows the official driver policy	Live query, mutation, and rollback baseline; migration compiler/unit verified
Neo4j GDS	GDS 2.13 is the official match for Neo4j 5.26	GDS must match the server according to Neo4j's compatibility matrix	Current procedure compiler verified; per-algorithm 2.13 live certification pending
TigerGraph	TigerGraph 4.2.4 Community	No wider Nodal compatibility promise yet	Live REST++ mutation/query baseline; analytics endpoint contract verified

References: Neo4j .NET Driver compatibility and Neo4j–GDS compatibility matrix.

The Docker images in compose.local.yml are the executable source for the live QA versions. Supporting a newer database version requires updating that baseline and passing the live suite; changing only this table is not sufficient.

Dependabot checks NuGet, documentation npm packages, Docker database images, and GitHub Actions every week. Its pull requests target developer. An update PR is a notification, not an automatic compatibility promise: database baselines move only after the relevant compiler and live integration suites pass.

Provider feature matrix

Legend: Yes is implemented; Conditional requires the condition shown; No fails explicitly.

Capability	Neo4j	TigerGraph
Parameterized node filtering and ordering	Yes, Cypher	Yes, interpreted GSQL
Fixed directed/undirected traversal	Yes	Yes
Variable-depth traversal	Yes	Conditional: GSQL Syntax V2
Optional match	Yes	No
Vertex-simple variable-depth path	Yes	No: intermediate aliases are unavailable
Client-managed multi-command transaction	Yes	No: request or installed-query boundary
Atomic mutation plan	Yes	Conditional: REST++ or installed mutation query
Migration execution	Yes	Conditional: administrative transport
Centrality/community analytics	Conditional: compatible GDS and named projection	Conditional: configured installed GSQL query
Weighted analytics	Algorithm-specific GDS capability	Explicitly declared per installed query
Analytics deployment discovery	Live GDS discovery with bounded cache	Configured installed-query snapshot
Analytics projection lifecycle	Yes: explicit create/reuse/drop	Not applicable; algorithms use the database graph
Typed shortest-path result	Yes: native Cypher	Conditional: configured installed GSQL query
Weighted shortest paths	Conditional: GDS Dijkstra, A*, Yen	Conditional: configured weighted installed query

Analytics algorithm matrix

All algorithms below exist in the portable contract. “Compiler” does not mean that the optional database component has been installed or live-certified.

Family	Algorithms	Neo4j	TigerGraph
Centrality	ArticleRank, articulation points, betweenness, bridges, CELF, closeness, degree, eigenvector, harmonic, HITS, PageRank	GDS stream compiler; deployment allow-list supported	Installed-query mapping; unavailable until configured
Community and cohesion	Clique counting, conductance, HDBSCAN, K-core, K-1 coloring, K-means, label propagation, Leiden, local clustering coefficient, Louvain, modularity, modularity optimization, SCC, triangle count, WCC, approximate maximum k-cut, SLLPA	GDS stream compiler; deployment allow-list supported	Installed-query mapping; unavailable until configured
Path finding	Unweighted shortest path, all shortest paths, Dijkstra, A*, Yen k-shortest paths	Native Cypher for unweighted paths; GDS compiler for weighted paths	Installed-query mapping with canonical route response

Neo4j can restrict advertised algorithms to the procedures actually available in a deployment:

var options = new Neo4jOptions
{
    Endpoint = new Uri("neo4j://localhost:7687"),
    Username = "neo4j",
    Password = "secret",
    GraphDataScienceEnabled = true,
    AnalyticsAlgorithms = new HashSet<GraphAnalyticsAlgorithm>
    {
        GraphAnalyticsAlgorithm.PageRank,
        GraphAnalyticsAlgorithm.Louvain
    }
};

Use context.Database.GetAnalyticsRuntime().DiscoverAsync() to read and cache gds.list(), gds.version(), and projection names. The deployment allow-list remains explicit so a newly installed server procedure cannot silently expand application permissions.

TigerGraph requires both the installed-query mapping and an explicit declaration for queries that accept weights:

var options = new TigerGraphOptions
{
    Endpoint = new Uri("https://example.i.tgcloud.io/"),
    AccessToken = "secret-token",
    AnalyticsQueries = new Dictionary<GraphAnalyticsAlgorithm, string>
    {
        [GraphAnalyticsAlgorithm.PageRank] = "nodal_pagerank"
    },
    WeightedAnalyticsAlgorithms = new HashSet<GraphAnalyticsAlgorithm>
    {
        GraphAnalyticsAlgorithm.PageRank
    }
};

An unsupported algorithm or semantic option is rejected before database transport. Nodal never downloads the graph to simulate missing server functionality.

Version policy​

Provider feature matrix​

Analytics algorithm matrix​

Version policy

Provider feature matrix

Analytics algorithm matrix