---

name: snowflake-geospatial description: "Use when working with geospatial data in Snowflake, including GEOGRAPHY columns, WKT or GeoJSON loading, spatial joins, ST_INTERSECTS or ST_DISTANCE queries, geohash workflows, and performance-aware SQL design for geo analytics in Snowflake."

Snowflake Geospatial

Use this skill for geospatial workflows that specifically target Snowflake.

Goals

• Choose Snowflake geospatial patterns that are correct and practical.
• Help load, transform, and query geospatial data using Snowflake SQL.
• Avoid common mistakes around types, CRS assumptions, and expensive spatial joins.
• Produce query plans that are explicit about correctness and cost.

Workflow

1. Clarify the storage shape.
   • Is the source data arriving as GeoJSON, WKT, WKB, coordinate columns, or an upstream table?
   • Identify the target geospatial column and whether the data should be stored as GEOGRAPHY.
2. Confirm spatial semantics.
   • Snowflake geospatial work is commonly centered on GEOGRAPHY with longitude and latitude semantics.
   • If the task needs projected analysis or strict planar behavior, state that Snowflake may need preprocessing outside the warehouse.
3. Design the SQL path.
   • Loading and parsing.
   • Cleaning and normalizing.
   • Spatial predicate or measurement.
   • Output summarization.
4. Guard the query for scale.
   • Reduce rows before expensive spatial predicates.
   • Prefer bounding or prefilter logic when possible.
   • Avoid unnecessary repeated geospatial construction inside joins.
5. Validate results.
   • Spot-check counts.
   • Sanity-check known distances or containment cases.
   • Confirm coordinate ordering and null handling.

Snowflake-Specific Rules

• Be explicit when creating GEOGRAPHY values from raw text or JSON.
• Treat axis order carefully: longitude first, then latitude, unless the source format guarantees otherwise.
• Do not assume Snowflake is the right place for every projection-heavy workflow.
• Separate correctness guidance from performance guidance.
• When writing SQL, prefer readable CTEs over dense single-block queries for spatial logic.

Common Tasks

• Load GeoJSON or WKT into a Snowflake table.
• Build point geometries from longitude and latitude columns.
• Run point-in-polygon joins.
• Filter records by distance from a point.
• Aggregate events by administrative boundary.
• Prepare geospatial outputs for BI or downstream analytics.

Output Format

Snowflake geospatial objective:
Recommended data type and storage pattern:
Key SQL approach:
Performance considerations:
Validation checks:
Risks / assumptions:

Example Triggers

• "Write a Snowflake query to join points to polygons"
• "How should I store GeoJSON in Snowflake for geo analytics?"
• "Help optimize this Snowflake spatial join"
• "Use Snowflake geography functions to filter by distance"