Field and type options

Available field types and options.

Redis Stack provides various field types that allow you to store and search different kinds of data in your indexes. This page explains the available field types, their characteristics, and how they can be used effectively.

Number fields

Number fields are used to store non-textual, countable values. They can hold integer or floating-point values. Number fields are sortable, meaning you can perform range-based queries and retrieve documents based on specific numeric conditions. For example, you can search for documents with a price between a certain range or retrieve documents with a specific rating value.

You can add number fields to a schema in FT.CREATE using this syntax:

FT.CREATE ... SCHEMA ... {field_name} NUMBER [SORTABLE] [NOINDEX]

where:

SORTABLE indicates that the field can be sorted. This is useful for performing range queries and sorting search results based on numeric values.
NOINDEX indicates that the field is not indexed. This is useful for storing numeric values that you don't want to search for, but that you want to retrieve in search results.

You can search for documents with specific numeric values using the @<field_name>:[<min> <max>] query syntax. For example, this query finds documents with a price between 200 and 300:

FT.SEARCH products "@price:[200 300]"

You can also use the following query syntax to perform more complex numeric queries:

Comparison operator	Query string	Comment
min <= x <= max	@field:[min max]	Fully inclusive range
min < x < max	@field:[(min (max]	Fully exclusive range
x >= min	@field:[min +inf]	Upper open range
x <= max	@field:[-inf max]	Lower open range
x == val	@field:[val val]	Equal
x != val	-@field:[val val]	Not equal

Geo fields

Geo fields are used to store geographical coordinates such as longitude and latitude. They enable geospatial radius queries, which allow you to implement location-based search functionality in your applications such as finding nearby restaurants, stores, or any other points of interest.

You can add geo fields to the schema in FT.CREATE using this syntax:

FT.CREATE ... SCHEMA ... {field_name} GEO [SORTABLE] [NOINDEX]

Where:

SORTABLE indicates that the field can be sorted. This is useful for performing range queries and sorting search results based on coordinates.
NOINDEX indicates that the field is not indexed. This is useful for storing coordinates that you don't want to search for, but that you still want to retrieve in search results.

You can query geo fields using the @<field_name>:[<lon> <lat> <radius> <unit>] query syntax. For example, this query finds documents within 1000 kilometers from the point 2.34, 48.86:

FT.SEARCH cities "@coords:[2.34 48.86 1000 km]"

Vector fields

Vector fields are floating-point vectors that are typically generated by external machine learning models. These vectors represent unstructured data such as text, images, or other complex features. Redis Stack allows you to search for similar vectors using vector similarity search algorithms like cosine similarity, Euclidean distance, and inner product. This enables you to build advanced search applications, recommendation systems, or content similarity analysis.

You can add vector fields to the schema in FT.CREATE using this syntax:

FT.CREATE ... SCHEMA ... {field_name} VECTOR {algorithm} {count} [{attribute_name} {attribute_value} ...]

Where:

{algorithm} must be specified and be a supported vector similarity index algorithm. The supported algorithms are:
- FLAT: brute force algorithm.
- HNSW: hierarchical, navigable, small world algorithm.
The {algorithm} attribute specifies the algorithm to use when searching k most similar vectors in the index or filtering vectors by range.
{count} specifies the number of attributes for the index and it must be present. Notice that {count} represents the total number of attribute pairs passed in the command. Algorithm parameters should be submitted as named arguments.

For example:
```
FT.CREATE my_idx SCHEMA vec_field VECTOR FLAT 6 TYPE FLOAT32 DIM 128 DISTANCE_METRIC L2
```
Here, three parameters are passed for the index (TYPE, DIM, DISTANCE_METRIC), and count is the total number of attributes (6).
{attribute_name} {attribute_value} are algorithm attributes for the creation of the vector index. Every algorithm has its own mandatory and optional attributes.

For more information about vector fields, see vector fields.

Tag fields

Tag fields are used to store textual data that represents a collection of data tags or labels. Tag fields are characterized by their low cardinality, meaning they typically have a limited number of distinct values. Unlike text fields, tag fields are stored as-is without tokenization or stemming. They are useful for organizing and categorizing data, making it easier to filter and retrieve documents based on specific tags.

Tag fields can be added to the schema with the following syntax:

FT.CREATE ... SCHEMA ... {field_name} TAG [SEPARATOR {sep}] [CASESENSITIVE]

where

SEPARATOR defaults to a comma (,), and can be any printable ASCII character. It is used to separate tags in the field value. For example, if the field value is hello,world, the tags are hello and world.
CASESENSITIVE indicates that the field is case-sensitive. By default, tag fields are case-insensitive.

You can search for documents with specific tags using the @<field_name>:{<tag>} query syntax. For example, this query finds documents with the tag blue:

FT.SEARCH idx "@tags:{blue}"

For more information about tag fields, see Tag Fields.

Text fields

Text fields are specifically designed for storing human language text. When indexing text fields, Redis Stack performs several transformations to optimize search capabilities. The text is transformed to lowercase, allowing case-insensitive searches. The data is tokenized, meaning it is split into individual words or tokens, which enables efficient full-text search functionality. Text fields can be weighted to assign different levels of importance to specific fields during search operations. Additionally, text fields can be sorted based on their values, enabling the sorting of search results by relevance or other criteria.

Text fields can be added to the schema with the following syntax:

FT.CREATE ... SCHEMA ... {field_name} TEXT [WEIGHT] [NOSTEM] [PHONETIC {matcher}] [SORTABLE] [NOINDEX] [WITHSUFFIXTRIE]