bigquery-graph

Use the GRAPH clause followed by MATCH patterns with node and edge labels, then RETURN the results. BigQuery supports only the GoogleSQL GQL standard, never Cypher, and path variables should be assigned for traversal queries.

Use CREATE PROPERTY GRAPH DDL with NODE TABLES and EDGE TABLES blocks mapping existing BigQuery tables. Define KEY columns, SOURCE KEY and DESTINATION KEY references, and scope properties explicitly rather than using PROPERTIES ALL COLUMNS.

Use GRAPH_TABLE only when you need SQL integration such as aggregations, joins with relational tables, or advanced filtering. For pure graph pattern matching, prefer standalone GQL with native RETURN statements.

Measure columns cannot be selected directly and must be wrapped in the AGG() function. SELECT * also fails on graphs with measures, so explicitly list non-measure columns or use EXCEPT to exclude measures.

GRAPH_EXPAND requires a tree structure and does not support disconnected graphs, directed cycles, self-loops, or convergent diamond paths. For cyclic graphs, use standard GQL pattern matching instead of the flattened view.

Common causes include starting traversals from high-cardinality nodes, omitting explicit labels, and using bidirectional edge patterns. Filter early, specify labels, and replace undirected patterns with directional queries combined via UNION ALL.