session_user

Generic Description

Returns session, schema, graph, or null-handling metadata inside the query. It is documented separately because execution context, graph context, and null-handling helpers can change how assistant-generated queries are validated.

Simple example:

RETURN session_user()

Consumer-Level Explanation

Use it when the query should expose execution context or normalize nullable values explicitly. Prefer this function when the query needs explicit context or null behavior instead of relying on an implicit client default.

More Detailed Explanation

session_user should be understood by the value shape it expects and the row shape it returns. It composes with MATCH, WHERE, WITH, RETURN, and procedure output, so planner-facing documentation should describe both the immediate value transformation and why it belongs in the database query.

Advanced Example

This example uses session_user inside a named row pipeline so the value shape is clear to both the planner and reviewers.

MATCH (u:User)-[e:VIEWED]->(d:Document)
TIME e.ts BETWEEN datetime('2025-01-01T00:00:00Z') AND datetime('2025-02-01T00:00:00Z')
WITH u, e, d,
     unpivot(properties(d.metadata)) AS metadata_rows,
     date_trunc('day', e.ts) AS day_bucket,
     session_user(current_graph()) AS focused_value
RETURN u.user_id, d.title, day_bucket, metadata_rows, focused_value
LIMIT 10

Real Use Cases

• keeping value shaping visible to the planner and reviewers
• combining graph structure with document-style payload handling
• returning named, typed fields that downstream tooling can reason about

Real Limitations And Tradeoffs

• the function does not repair incorrectly modeled input values
• nearby aliases may be clearer if they express the business intent more directly
• broad use inside hot filters should be reviewed for cost and index interaction