Возможности геопространственного поиска

Welcome to Geospatial Search! Many modern applications rely on location data, from finding nearby restaurants to tracking delivery vehicles.

Elasticsearch provides powerful features to index, search, and analyze geographic information efficiently.

The most common way to store a single geographical location in Elasticsearch is using the geo_point data type.

It represents a specific point on Earth using its latitude and longitude coordinates.

• Latitude: North-South position (e.g., 48.8584)
• Longitude: East-West position (e.g., 2.2945)

Before indexing documents with location data, you need to define a geo_point field in your index mapping. This tells Elasticsearch how to interpret the coordinates.

Try setting up an index named places with a location field of type geo_point:

Now that our places index is mapped, we can index documents containing geo_point data. The location field can accept an object with lat and lon properties.

Index the Eiffel Tower's location:

One of the most common geospatial queries is finding documents within a certain radius of a central point. This is achieved using the geo_distance query.

You specify a distance and the reference point, and Elasticsearch returns all matching documents.

Let's find all places within 5 kilometers of the Eiffel Tower's coordinates. This query is perfect for 'find restaurants near me' features.

Run this query:

Another useful query is geo_bounding_box, which allows you to find all documents whose geo_point falls within a specified rectangular area.

You define the box by providing its top_left and bottom_right coordinates.

Imagine you have a map view and only want to show results within the current screen. A geo_bounding_box query is ideal for this scenario.

Here's an example searching within a small area around Paris:

While geo_point is great for single locations, what if you need to represent complex geometries like a park boundary, a river, or a country's outline?

For this, Elasticsearch offers the geo_shape data type, which supports various GeoJSON shapes (polygons, lines, multi-points, etc.).

To use geo_shape, you first map the field as geo_shape. Then, you can index documents with complex geometries, typically in GeoJSON format. Here's how to map and index a simple polygon representing a district:

You want to find all restaurants located within the boundaries of a specific park. The park's boundaries are complex and not a simple rectangle.

You've explored Elasticsearch's powerful geospatial capabilities!

• We learned about the geo_point type for single locations (latitude, longitude).
• We used geo_distance for proximity searches and geo_bounding_box for rectangular area searches.
• We also introduced the advanced geo_shape type for complex geometries like polygons.

These tools are essential for building sophisticated location-aware applications.