Grammar

Cypher Path Subset is defined as its own ANTLR parser grammar, derived from Neo4j’s Cypher 25 grammar. The railroad diagram below is generated directly from that grammar and shows every rule the parser accepts.

Open in new tab

How it differs from Cypher 25

Cypher Path Subset keeps only what is needed to describe a path pattern and return it - most of Cypher 25’s surface area is removed:

  • No write access: CREATE, INSERT, DELETE, SET, REMOVE and MERGE are all gone - a query can only read.
  • No function calls or existence checks: functionInvocation, existsExpression and countExpression are not part of the grammar, which rules out arbitrary computation and nested sub-queries inside the query itself.
  • No procedures or sub-queries: CALL (both standalone and as a subquery), LOAD CSV, FOREACH and USE are removed.
  • A single, non-chainable pipeline: a query is exactly one optional UNWIND followed by one MATCH followed by one RETURN - there’s no WITH clause to chain multiple stages, and no UNION to combine queries.
  • A fixed, minimal RETURN: returnItem only allows a bare variable, not arbitrary expressions or AS aliases - the query can only return the variable paths (see step reference).

ORDER BY, SKIP and LIMIT on the result, as well as WHERE filters on patterns, are still fully supported - see the Cypher path query examples for both in use.

See also