GraphQuery

A flat aggregation over the anchor node set: buckets nodes by the groupBy fields and computes the requested metrics per bucket. Supported only for resource anchors, and not combinable with traversal.

Selects the anchor (zero-hop) node set. 'match' and 'query' are mutually exclusive, and supplying neither selects every node of the given type that the caller can see.

Pagination controls. Results are paged with an opaque keyset token: read pageInfo.continuationToken from a response and pass it back as 'continuationToken' to fetch the next page.

Shapes the returned subgraph.

Bounds anchor selection to a subset of the organization. Scope restricts only where anchors are selected, never where traversal may reach.

Ordered, typed, depth-bounded traversal steps applied from the anchor set. Each step's result becomes the frontier for the next step.

Aggregation results: one bucket per distinct groupBy key combination.

The typed edges revealed by this page: every edge whose higher-id endpoint first appears on this page. Both endpoints are among the nodes returned by this page or an earlier one — process pages cumulatively. A drain over changing data may not be complete.

Metadata about how a graph query was answered.

This page's slice of the selected nodes (id-ordered), from the frontiers named in return.select plus any nodes referenced by returned paths.

Result-set pagination information. A present continuationToken means more pages remain, whereas meta.resultMode=truncated means an engine cap or a partial search-backend answer clipped the underlying result.

Evidence paths showing how the query reached its results. Present only when return.paths is true. One path per node of the final traversal frontier, not necessarily the shortest route.