Schemas define the data structure to be used.
They allow for data to be automatically validated and the right type of index to be used for searching.
These fields are added automatically as part of each type.
name: stringchildren: referencesparents: references
| Name | type | atributes | description |
|---|---|---|---|
languages |
Array | optional | Sets the available languages in the data |
types |
Object | Defines the types. Each object key is the type name and it's value the type definition. |
{
//...
types: {
league: { // type name
prefix: 'le', // type prefix
fields: {
name: { type: 'string', search: { type: ['TAG'] } }
}
}
}
}| Name | type | atributes | description |
|---|---|---|---|
prefix |
string | optional | Two character string that identifies the type. Used as a prefix for each object/id |
fields |
Object | Defines the fields for the type. Each object key is a field name and its value the field difinition. |
Defines the fields available to each type.
| Name | type | atributes | description |
|---|---|---|---|
type |
string | Two character string that identifies the type. Used as a prefix for each object/id | |
fields |
Object | Data type for the field. | |
search |
Object | optional | Index type for the field. |
Object allow for nested structures, but unlike the JSON type, they allow its properties to have type validation and indexes.
In field of type object, the subfield properties is a reserved field name for this feature.
{
// ...
movie: {
prefix: 'mo',
fields: {
title: { type: 'text' },
technicalData: {
type: 'object',
properties: {
runtime: { type: 'int' },
color: { type: 'string' },
aspectRatio: { type: 'string' }
}
}
}
}
}Field indexes are described in the schema objects of the following format:
{
index: 'default' // a `string` value, optional, defaults to the `default` index
type: ['TAG', 'EXISTS'] // array of field index types
}
See index types below for more information.
For the fields to be searchable, it needs an index. There are several types of indexes avaliable. One field may have multiple index types.
The following combinations are supported:
['TAG']['TAG', 'EXISTS']['NUMERIC']['NUMERIC', 'EXISTS']['NUMERIC', 'SORTABLE']['NUMERIC', 'SORTABLE', 'EXISTS']['GEO']['GEO', 'EXISTS']['TEXT-LANGUAGE']['TEXT-LANGUAGE', 'EXISTS']['TEXT-LANGUAGE-SUG']['TEXT-LANGUAGE-SUG', 'EXISTS']
Below you will find a description for all the existing index types.
The TAG index type is most commonly applied to values of the following field types:
- string
- digest
- boolean
- url
- phone
but can also be applied to the following types:
- number
- float
- int
- timestamp
The type field is automatically indexed as a TAG.
When indexed as TAG, the field becomes queryable with $filter using the following operators:
$operator: '='$operator: '!='
The filter will match the field values as an exact match of the values.
The NUMERIC index type is most commonly applied to values of the following field types:
- number
- float
- int
but can also be applied to the following types:
- timestamp
When indexed as NUMERIC, the field becomes queryable with $filter using the following operators:
$operator: '='(exact equality of the numeric values)$operator: '!='(exact inequality of the numeric values)$operator: '>'(larger than$value)$operator: '<'(smaller than$value)$operator: '..'(is in the range from$value[0]to$value[1])
The SORTABLE index type is most commonly applied to values of the following field types (indexed as NUMERIC):
- number
- float
- int
- timestamp
but can also be applied to the following types:
- string (as
TAG) - text (as
TEXT-LANGUAGE)
When indexed as SORTABLE, the field becomes queryable with $sort.
The TEXT-LANGUAGE index type can be applied to the text type only.
When indexed as TEXT-LANGUAGE, the field becomes queryable with $filter using the following operators:
$operator: '='$operator: '!='
The text match will be a text search approximate string match, including preprocessing such as:
- Removing special characters
- Word stemming
- Excluding most common words if they don't affect the search results
Some additional operators are possible in the $value filter property:
$value: 'baye*' // would match bayern
$value: '%heylo%' // would match "hello", with one character substitution from "hello"
$value: '%%heyyo%%' // would match "hello", with two character substitutions from "hello"
The TEXT-LANGUAGE-SUG index type can be applied to the text type only.
When indexed as TEXT-LANGUAGE-SUG, the field becomes queryable with $filter using the following operators:
$operator: '='$operator: '!='
Matching with TEXT-LANGUAGE-SUG works much the same way as with TEXT-LANGUAGE, however, the $value property is first passed through a search autocomplete index. It is mostly used for things such as implementing the search bar for movie titles.
No operators are possible in the $value filter property itself, unlike with TEXT-LANGUAGE.
The GEO index type can be applied to the geo type only.
When indexed as GEO, the field becomes queryable with $filter using the following operator:
$operator: 'distance'
The distance operator can be used to return only results within a certain radius of a lat/lon location.
The EXISTS index type can be applied to any type. When other indexing is enabled, it can be added as the last index type. It enables the following $filter operators:
$operator: 'exists'(filter matches only records where the field has been set)$operator: 'notExists'(filter matches only records where the field has not been set)
floatbooleannumberintreferencesstringtextiddigesturlemailphonegeotimestampsetarrayobjectjson
For more information on what kind of values are stored in each of these types, you can visit the set query documentation.