Choosing the right data types and mappings in Elasticsearch

Elasticsearch is a highly scalable and powerful search engine that allows you to index and search large amounts of data quickly. To take full advantage of Elasticsearch's capabilities, it is crucial to choose the right data types and mappings for your documents.

Understanding data types

In Elasticsearch, data types are used to determine how data is indexed and stored, which affects the overall performance and behavior of your searches. Elasticsearch provides a variety of built-in data types, including:

  • Numeric types: Integers (long, integer, short, byte) and floating-point numbers (double, float).
  • Text types: Strings (text, keyword) used for full-text search or exact matching.
  • Date types: Dates (date) that can be indexed and queried using various formats.
  • Boolean types: True or false values (boolean).
  • Binary types: Binary data (binary), such as images or serialized objects.
  • Geo types: Geospatial data (geo_point, geo_shape) for indexing and querying points or shapes.

Choosing the right data type

Selecting the appropriate data type for your field is essential to ensure accurate and efficient searches. Here are a few guidelines to help you make the right choice:

Numeric types

  • If you need to perform mathematical calculations or range queries on the field, choose a numeric type.
  • Use integer types for whole numbers, and floating-point types for decimal numbers.
  • Be mindful of the range limits of each numeric type, as they affect storage size and precision.

Text types

  • Use text for full-text search fields, where you want to analyze and tokenize the content.
  • Use keyword for fields that require exact matching, such as IDs or categories.
  • Set appropriate analyzers to define how the text should be processed during indexing and searching.

Date types

  • Use date for fields containing date/time information.
  • Specify a date format that matches your data to ensure accurate indexing and searching.
  • Consider enabling the date_nanos format for nanosecond precision if required.

Boolean types

  • If your field can have only two states (true/false), use the boolean type for efficient storage and querying.

Binary types

  • Use the binary type for storing raw binary data, such as images.
  • Consider using dedicated tools or frameworks to handle binary data efficiently.

Geo types

  • For points on the globe, use geo_point type to store latitude and longitude information.
  • For more complex shapes, such as polygons or lines, use geo_shape type.
  • Ensure that your geo points or shapes are properly formatted for accurate indexing and querying.

Configuring mappings

Mappings in Elasticsearch define how your data is indexed and how fields are parsed. By default, Elasticsearch automatically detects the data type of your fields. However, explicitly defining the mappings can enhance search performance and provide better control over your data. When configuring mappings:

  • Specify the data type of each field explicitly to avoid unexpected behavior.
  • Define custom analyzers for text fields to control tokenization and normalization.
  • Set appropriate formats for date fields to handle different date/time representations.
  • Consider using multi-fields for a single field to support different search requirements.

Remember, mappings cannot be changed once data is indexed. Therefore, it is essential to plan and design your mappings accurately before indexing your documents.


Choosing the right data types and mappings in Elasticsearch plays a vital role in ensuring efficient search performance and accurate results. Understanding the purpose of your fields and the available data types will help you make informed decisions when configuring your mappings. With the proper choices, you can unleash the true power of Elasticsearch and deliver exceptional search experiences.

noob to master © copyleft