You are browsing a version that is no longer maintained. |
Search Indexes
In addition to standard indexes, ODM allows you to define search indexes for use with MongoDB Atlas Search. Search indexes may be queried using the $search and $searchMeta aggregation pipeline stages.
Search indexes have some notable differences from regular indexes in ODM. They may only be defined on document classes. Definitions will not be incorporated from embedded documents. Additionally, ODM will NOT translate field names in search index definitions. Database field names must be used instead of mapped field names (i.e. PHP property names).
Search Index Options
Search indexes are defined using a more complex syntax than regular indexes.
ODM supports the following search index options:
name
- Name of the search index to create, which must be unique to the collection. Defaults to"default"
.dynamic
- Enables or disables dynamic field mapping for this index. Iftrue
, the index will include all fields with supported data types. Iffalse
, thefields
attribute must be specified. Defaults tofalse
.fields
- Associative array of field mappings that specify the fields to index (keys). Required only if dynamic mapping is disabled.analyzer
- Specifies the analyzer to apply to string fields when indexing. Defaults to the standard analyzer.searchAnalyzer
- Specifies the analyzer to apply to query text before the text is searched. Defaults to theanalyzer
attribute, or the standard analyzer. if both are unspecified.analyzers
- Array of custom analyzers to use in this index.storedSource
- Specifies document fields to store for queries performed using the returnedStoredSource option. Specifytrue
to store all fields,false
to store no fields, or a document to specify individual fields to include or exclude from storage. Defaults tofalse
.synonyms
- Array of synonym mapping definitions to use in this index.
Additional documentation for defining search indexes may be found in search index definition within the MongoDB manual.
Static Mapping
Static mapping can be used to configure indexing of specific fields within a document.
The following example demonstrates how to define a search index using static mapping.
1 <?php
/**
* @Document
* @SearchIndex(
* name="usernameAndAddresses",
* fields={
* "username"={
* {"type"="string"},
* {"type"="autocomplete"},
* },
* "addresses"={"type"="embeddedDocuments", "dynamic"=true},
* },
* )
*/
class User
{
/** @Id */
private $id;
/** @Field(type="string") */
private $username;
/** @EmbedMany(targetDocument=Address::class) */
private $addresses;
// ...
}
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
The username
field will indexed both as a string and for autocompletion.
Since the addresses
field uses an embed-many
relationship, it must be indexed using the embeddedDocuments
type; however,
embedded documents within the array are permitted to use dynamic mapping.
Dynamic Mapping
Dynamic mapping can be used to automatically index fields with supported data types within a document. Dynamically mapped indexes occupy more disk space than statically mapped indexes and may be less performant; however, they may be useful if your schema changes or for when experimenting with Atlas Search
Atlas Search does NOT dynamically index embedded documents contained within arrays (e.g. embed-many relationships). You must use static mappings with the embeddedDocument field type. |
The following example demonstrates how to define a search index using dynamic mapping: