Generate DDL

rudof ddl derives a property graph schema from RDF data and generates the DDL needed to materialize it in a property graph database. It is stateless: it never opens a database, which makes it useful to inspect the schema before loading anything.

rudof ddl <data...> --dialect <cypher|gql>

Also works inside rudof shell, the same as any other subcommand.

The schema is discovered from the data:

• Every class used in an rdf:type triple becomes a node table (one column per predicate, plus an id primary key column holding the subject IRI).
• Every predicate whose object is an IRI referring to a typed node becomes a relationship table.

Dialects

--dialect cypher

LadybugDB/Kùzu-style Cypher DDL:

$ rudof ddl examples/user.ttl --dialect cypher
CREATE NODE TABLE Person (id STRING, knows STRING, name STRING, status STRING, PRIMARY KEY(id));
CREATE REL TABLE knows (FROM Person TO Person);

These are exactly the statements that rudof load applies to the database before inserting data.

--dialect gql

ISO GQL-style graph type DDL (CREATE GRAPH TYPE with NODE TYPE/EDGE TYPE declarations), which documents the graph type in the dialect of the GQL standard:

$ rudof ddl examples/user.ttl --dialect gql
CREATE GRAPH TYPE rudof_graph (
  NODE TYPE Person (id STRING, knows STRING, name STRING, status STRING),
  EDGE TYPE knows (FROM Person TO Person)
);

The name of the graph type can be customized with --graph-type-name <NAME>.

Options

Progress messages are written to stderr, so the generated DDL on stdout can be piped directly to another tool:

rudof ddl data.ttl --dialect cypher 2>/dev/null | cypher-shell

Relationship with pgschema

rudof already supports PGSchema for validating YARS-PG property graph data with rudof pgschema-validate. The ddl command is the DDL-emitting counterpart of that infrastructure: both operate on property graph schemas (node types with properties, edge types with endpoints) and differ only in serialization target — pgschema serializes to the PGSchema-C format, while ddl serializes to executable Cypher/GQL DDL.

The internal schema model produced by ddl is deliberately dialect-agnostic so that a future bridge can feed a PGSchema schema (.pgs file) directly into the DDL emitters, e.g. rudof ddl --schema user.pgs --dialect cypher, and conversely emit a PGSchema-C schema from RDF data — bridging RDF/SHACL semantics with storage-oriented graph technologies like Cypher and GQL.

The other direction of that bridge is --backend lbug, already accepted on RDF-loading commands (data, shacl-validate, …) alongside memory/qlever/endpoint=... (see The --backend flag), but reading a connected LadybugDB graph back out as RDF for SHACL validation isn’t implemented yet — selecting it there currently fails with a clear, specific error instead of silently doing nothing.