Query Benchmark — 100 Queries (Single Table)

Complexity Classification

• Simple (S): Single table or single join. No spatial functions, no aggregation beyond COUNT.
• Medium (M): Property join with namespace awareness, codelist resolution, basic aggregation, or one spatial function.
• High (H): 1-hop boundary surface relationship, 3D geometry computation, or multi-condition filtering with aggregation.
• Very High (VH): 2-hop chains, cross-module spatial joins, CTE arithmetic, or combination of 3D computation with multi-hop relationships.

Query Categories

• COUNT: Feature counting and listing
• PROP: Property access (schema properties, nested types)
• CODE: Codelist resolution and filtering
• AGG: Aggregation (SUM, AVG, GROUP BY)
• GEOM: 3D geometry computation (volume, area)
• SPAT: Spatial queries (distance, within, intersects)
• REL-1: 1-hop boundary surface relationship
• REL-2: 2-hop relationship chain
• CTE: CTE arithmetic (subtract, compare)
• GATTR: Generic attribute query
• CROSS: Cross-module query

Datasets

• H: Harthof District in Munich (DE, LoD 0–2, Buildings)