Skip to content
CSA Loom — the Microsoft Fabric experience for Azure tenants where Fabric isn't yet available: lakehouses, warehouses, notebooks, semantic models, Activator rules, Data Agents, across Commercial, GCC, GCC-High, and DoD IL5

Loom Geo Query Editor — Azure-native parity spec

Reference: Azure Data Explorer KQL geospatial functions (geo_* + S2/H3/Geohash cell systems) + Synapse Serverless geography/geometry SQL CLR types + SQL Server 2025 spatial indexes, captured 2026-05-26 by the catalog agent. There is no Fabric equivalent — geo-query is a Loom-native item that targets either an ADX/KQL backend or a Synapse Serverless backend depending on the linked geo-dataset.

What's there in the Azure-native experience

Kusto / ADX query surface

  • Built-in geo functions (all ✅ Microsoft Fabric ✅ Azure Data Explorer ✅ Azure Monitor ✅ Microsoft Sentinel):
  • Distance: geo_distance_2points, geo_distance_point_to_line, geo_distance_point_to_polygon
  • Containment: geo_point_in_polygon, geo_point_in_circle
  • Cell systems: geo_point_to_s2cell / geo_point_to_h3cell / geo_point_to_geohash + their *_to_polygon and *_to_central_point inverses
  • Covering: geo_polygon_to_s2cells, geo_line_to_s2cells, geo_polygon_to_h3cells
  • Buffers: geo_point_buffer, geo_line_buffer, geo_polygon_buffer
  • Lookup plugins: geo_polygon_lookup, geo_line_lookup (efficient classify-many-points-into-many-polygons join)
  • Densify / simplify: geo_polygon_densify, geo_polygon_simplify
  • Spatial joins via cell tokens — documented pattern: cover polygons + points with the same S2/H3 cell level, join on the cell, then filter with geo_point_in_polygon for exact membership. Recommended level: 5 for countries, 11 for suburbs, 16 for dense neighborhoods.

Synapse Serverless / SQL spatial surface

  • T-SQL geometry and geography CLR types with STWithin, STIntersects, STContains, STDistance, STOverlaps, STTouches, STEquals, STBuffer, STCentroid, STArea, STLength
  • Spatial index support on geometry methods within WHERE / JOIN ON predicates of the form geom1.STMethod(geom2) = 1 or geom1.STDistance(geom2) < N
  • SRID enforcement: two geometries with different SRIDs return NULL — must STSrid match or STTransform

Azure Maps query support

  • Search Inside Geometry REST POST /search/inside/json — find POIs inside an arbitrary polygon
  • Search Along Route REST POST /search/alongRoute/json — find POIs within a corridor of a polyline

Databricks SQL geospatial (Public Preview, DBR 17.1+)

  • 30+ st_* functions including st_within, st_intersects, st_dwithin, st_buffer, st_distance, st_transform, st_setsrid, st_estimatesrid

What Loom already has

  • ✅ GeoQueryEditor stub with KQL / T-SQL toggle
  • ✅ Sample KQL query pre-populated with geo_distance_2points + geo_point_to_h3cell
  • ✅ Sample T-SQL query pre-populated with OPENROWSET against ADLS Parquet + GEOGRAPHY::STGeomFromText + STDistance
  • ✅ Submit path connects to existing /api/items/synapse-serverless-sql-pool/[id]/query and Kusto routes
  • ✅ Honest MessageBar noting that H3 SQL UDFs aren't deployed in Serverless by default

What's missing for parity

  1. Backend auto-select — based on the linked geo-dataset, automatically pick KQL (when source is ADX/Eventhouse) or T-SQL (when source is ADLS Parquet/GeoParquet)
  2. Spatial function autocomplete — Monaco editor with a curated KQL geo_* and T-SQL ST* snippet library + parameter hints
  3. Visual query builder — form mode for the common patterns: "find features within radius of point", "find features intersecting polygon", "classify points into polygons", "compute distance between two columns"
  4. Result map preview — when result has a geometry column, render results on an inline Azure Maps preview (uses geo-map rendering helper)
  5. Geometry input picker — instead of typing GeoJSON inline, draw a polygon/point on a mini-map and bind it to a @geomParam parameter
  6. H3 / S2 UDF install action — one-click "Install H3 UDFs into Synapse Serverless" button that runs the CREATE FUNCTION script pointing at the H3 wheel in the lake
  7. Cell-cover join optimizer — when the user writes a JOIN ... ON ST_Within(point, poly) pattern, surface a "Convert to S2 cell join" rewrite suggestion (matches the Kusto docs' recommended performance pattern)
  8. Saved-query library — Cosmos-backed list of saved geo queries per workspace; "Open in geo-pipeline" promotes a saved query into a scheduled job
  9. Spatial index hints — when targeting Azure SQL/MI, surface WITH(INDEX(spatial_idx)) hint controls
  10. Result export to geo-dataset — "Save result as new dataset" writes the result back to ADLS as GeoJSON/GeoParquet

Backend mapping

Loom surface Real Azure call
KQL submit Existing Kusto POST https://{cluster}.{region}.kusto.windows.net/v2/rest/query via Loom's Eventhouse editor route
T-SQL submit Existing Synapse Serverless TDS path via /api/items/synapse-serverless-sql-pool/[id]/query
Azure SQL/MI submit /api/items/azure-sql-database/[id]/query
Search Inside Geometry POST https://atlas.microsoft.com/search/inside/json?api-version=1.0&query=<poi>
Install H3 UDF Spawn Synapse Serverless CREATE FUNCTION dbo.H3_LATLON_TO_CELL(...) RETURNS bigint AS EXTERNAL NAME ... pointing at the H3 .NET assembly in the lake
Map preview of result Reuse geo-map Azure Maps Web SDK canvas in a side panel

Required Azure resources

  • One or more of: Azure Data Explorer cluster / Synapse Serverless SQL pool / Azure SQL DB / SQL MI (Loom has wiring for all of them)
  • ADLS Gen2 container for GeoParquet inputs (existing)
  • (optional) Microsoft.Maps/accounts for Search Inside Geometry + Search Along Route REST calls
  • (optional) H3 .NET assembly published to ADLS for the Serverless H3 UDF install path
  • Role: managed identity needs Database Reader on Kusto, db_datareader on Serverless, Storage Blob Data Reader on ADLS

Build plan

Phase Work
Backend New /api/items/geo-query/[id]/run → routes to the correct backend (Kusto / Serverless / SQL DB) based on linked dataset; returns rows + geometry GeoJSON. New /api/items/geo-query/[id]/install-h3-udfs (POST) → emits the CREATE FUNCTION script. New /api/items/geo-query/[id]/save-as-dataset (POST) → writes result back as GeoParquet.
Frontend Monaco editor with geo_* and ST* IntelliSense; tab switcher Code / Builder; side panel "Map preview"; toolbar with Run · Save · Save as dataset · Install H3 UDFs · Convert to S2-join.
Templates Ship a starter library of 10 saved queries: point-in-polygon · radius search · H3 hotspot · S2 join · polygon area · centroid · buffer · route corridor search · isochrone reachability · multi-polygon classify.

Estimated effort

2 focused sessions. Backend routing + result map preview in session 1, Monaco IntelliSense + visual builder + H3 install action in session 2.