Manage indexes

MongoDB API

Use the createIndex() method to create an index. For example:

•   
  db.restaurants.createIndex ({ 
   "cuisine" 
    
  :  
   1 
   }) 
    
  

•   
  db.restaurants.createIndex ({ 
   "cuisine" 
    
  :  
   1 
   } 
  ,  
   { 
  sparse:  
  true }) 
    
  

• Index creation with db.runCommand() is also supported with at most one index.

    
  db.runCommand ({ 
   "createIndexes" 
  : "restaurant" 
  ,  
   "index" 
  :  
   [{ 
   "key" 
  :  
   { 
   "cuisine" 
  :1 } 
  ,  
   { 
   "name" 
  :  
   "cuisine_index" 
   }]}) 
    
  

Note the following limitations:

• You can create only one index per request. db. collection .createIndexes() is not supported.
• Audit logs for index creation with the MongoDB API use the method name google.firestore.admin.v1.FirestoreAdmin.CreateIndex .
• For supported index options, see indexes and index properties .

Firebase console

1. In the Firebase console, go to the Firestore Database page.

   Go to Firestore Database

2. Select a database from the list of databases.
3. In the Indexes tab, click Create Index .
4. Enter a Collection ID .
5. Add one or more field paths and select an index option for each.
6. Select a field presence option, either non-sparse or sparse.
7. Optionally, you can set the multikey index option.
8. Click Create .
9. Your new index is displayed in the list of indexes and Cloud Firestore begins creating your index. When your index is created, you will see a green check mark next to the index. If index is not created, see Index building errors for possible causes.

gcloud CLI

To create an index, use the gcloud firestore indexes composite create command. Set api-scope to mongodb-compatible-api .

gcloud firestore indexes composite create \
--database=' DATABASE_ID 
' \
--collection-group= COLLECTION 
\
--field-config= FIELD_CONFIGURATION 
\
--query-scope=collection-group \
--density=dense \
--api-scope=mongodb-compatible-api

Replace the following:

• DATABASE_ID : a database ID.
• COLLECTION : a collection name.
• FIELD_CONFIGURATION : a field configuration. For each field, add --field-config=field-path= . For example:

  --field-config=field-path=user-id,order=descending \
      --field-config=field-path=score,order=descending

  For more information about configuring these fields, see --field-config .

To create a sparse index, set --density=sparse-any .

To create a multikey index, add the --multikey flag.

To create a unique index, add the --unique flag.

Terraform

Use the google_firestore_index resource and set api_scope to MONGODB_COMPATIBLE_API and query_scope to COLLECTION_GROUP .

 resource 
  
 "google_firestore_index" 
  
 "index" 
  
 { 
  
 database 
  
 = 
  
 " DATABASE_ID 
" 
  
 collection 
  
 = 
  
 " COLLECTION 
" 
  
 api_scope 
  
 = 
  
 "MONGODB_COMPATIBLE_API" 
  
 query_scope 
  
 = 
  
 "COLLECTION_GROUP" 
 // You can include multiple field blocks 
  
 fields 
  
 { 
  
 field_path 
  
 = 
  
 " FIELD_PATH 
" 
  
 order 
  
 = 
  
 " ORDER 
" 
  
 } 
 // Optional 
  
 multikey 
  
 = 
  
 true 
  
 density 
  
 = 
  
 " DENSITY 
" 
 } 

Replace the following:

• DATABASE_ID : The database ID for your chosen database
• COLLECTION : The name of the collection to index
• FIELD_PATH : The name of the field to index
• ORDER : One of ASCENDING or DESCENDING
• DENSITY : One of SPARSE_ANY or DENSE

MongoDB API

Use the createIndex() method to create a text index. In the following example, when a document is written to the cities collection with the country , or food fields populated, these fields are indexed for searching purposes.

 db.cities.createIndex ({ 
 "country" 
:  
 "text" 
,  
 "food" 
:  
 "text" 
 }) 
 

A field must be a string or an array of strings to be indexed. Array indexes are not search-indexed. As a result, indexing a.1.b will index something in {a: {1: {b: something}}} , but not in {a: [one, {b: something}]} .

You can also create indexes with db.runCommand() . You can have only one text index per collection, but you can create multiple indexes of different types in one db.runCommand() . The following example uses db.runCommand() to create a text index:

 db.runCommand ({ 
  
createIndexes:  
 "cities" 
,  
indexes:  
 [ 
  
 { 
  
key:  
 { 
  
 "country" 
:  
 "text" 
,  
 "food" 
:  
 "text" 
  
 } 
,  
name:  
 "country_text_food_text" 
  
 } 
  
 ] 
 }) 
 

Specify a default language

You can also optionally specify a default language, or a field path in your document that will contain the default language.

In the following example, myLanguageField is specified as the language_override . When a document in the cities collection contains a field named myLanguageField , the value of that field is used to determine the language for indexing the country field for that specific document. That value overrides the default language of french .

 db.cities.createIndex ({ 
 "country" 
:  
 "text" 
 } 
,  
 { 
 "default_language" 
:  
 "french" 
,  
 "language_override" 
:  
 "myLanguageField" 
 }) 
 

• You can enter a language as either its long form name ( english ) or its two-letter ISO language code ( en ).
• If the default language is left unset, then Cloud Firestore defaults to English.
• The language override field must be a top-level field. If left unset, the language override field defaults to language .
• If you set the default language to the null character, then Cloud Firestore doesn't use any field as a language override.

Partition a text index

You can also partition your index using a field so that you can filter queries by a specific field value. This configuration lets you run more peformant queries if you always need a specific field filtered in the index you're querying.

To create an index with a partition, configure the firestoreOptions field as follows:

 db.runCommand ({ 
  
createIndexes:  
 "cities" 
,  
indexes:  
 [ 
  
 { 
  
key:  
 { 
  
 "country" 
:  
 "text" 
,  
 "food" 
:  
 "text" 
 } 
,  
name:  
 "country_text_food_text" 
  
firestoreOptions:  
 { 
 "customPartitionFields" 
:  
 [ 
 " PARTITIONED_FIELD 
" 
 ] 
  
 } 
  
 ] 
 }) 
 

Where:

• PARTITIONED_FIELD is the name of the field used for the partition. This value must be a string and must refer to a top-level field. When you run a query against a partitioned index, you can filter the results based on a value of this field. For example, you could partition your index using city . If a city field is defined in your text index, then users can query against a specific city.

  The partition must be only one field. If you partition an index, then you can run only queries where the partitioned field is specified.

Limitations

• You can create only one index per request.
• Audit logs for index creation with the MongoDB API use the method name google.firestore.admin.v1.FirestoreAdmin.CreateIndex .
• For supported index options, see indexes and index properties .

Firebase console

1. In the Firebase console, go to the Firestore Databasepage.

   Go to Firestore Database

2. Select a database from the list of databases.

3. In the Indexestab, click Create Index.

4. Enter a Collection ID.

5. Add one or more field paths and select an index option for each.

6. Click Create.

7. Your new index is displayed in the list of indexes and MongoDB compatible operations begins creating your index. When your index is created, you will see a green check mark next to the index. If index is not created, see Index building errors for possible causes.

MongoDB API

Use the createIndex() method to create an index. For example:

 db.restaurants.createIndex ({ 
 "location" 
  
:  
 "2dsphere" 
,  
 "region" 
:  
 "2dsphere" 
 }) 
 

Index creation with db.runCommand() is also supported with at most one index:

 db.runCommand ({ 
  
createIndexes:  
 "restaurants" 
,  
indexes:  
 [ 
  
 { 
  
key:  
 { 
  
 "location" 
:  
 "2dsphere" 
,  
 "region" 
:  
 "2dsphere" 
  
 } 
,  
name:  
 "location_2dsphere_region_2dsphere" 
  
 } 
  
 ] 
 }) 
 

Partition a 2dsphere index

You can also partition your index using a field so that you can filter queries by a specific field value. This configuration lets you run more peformant queries if you always need a specific field filtered in the index you're querying.

To create an index with a partition, configure the firestoreOptions field as follows:

 db.runCommand ({ 
  
createIndexes:  
 "restaurants" 
,  
indexes:  
 [ 
  
 { 
  
key:  
 { 
  
 "location" 
:  
 "2dsphere" 
,  
 "region" 
:  
 "2dsphere" 
  
 } 
,  
name:  
 "location_2dsphere_region_2dsphere" 
  
firestoreOptions:  
 { 
 "customPartitionFields" 
:  
 [ 
 " PARTITIONED_FIELD 
" 
 ] 
  
 } 
  
 ] 
 }) 
 

Where:

• PARTITIONED_FIELD is the name of the field used for the partition. When you run a query against a partitioned index, you can filter the results based on a value of this field. For example, if your index has a region field for regional locations, you could partition your index using region so that users can query against restaurants in their region.

  If you partition an index, then you can only run queries where the partitioned field is specified.

Limitations

• You can create only one index per request.
• Audit logs for index creation with the MongoDB API use the method name google.firestore.admin.v1.FirestoreAdmin.CreateIndex .
• For supported index options, see indexes and index properties .

Firebase console

1. In the Firebase console, go to the Firestore Databasepage.

   Go to Firestore Database

2. Select a database from the list of databases.

3. In the Indexestab, click Create Index.

4. Enter a Collection ID.

5. Add one or more field paths and select an index option for each.

6. Click Create.

7. Your new index is displayed in the list of indexes and MongoDB compatible operations begins creating your index. When your index is created, you will see a green check mark next to the index. If index is not created, see Index building errors for possible causes.

MongoDB API

Use the dropIndex() method to delete an index. For example:

Delete an index using index name

db.restaurants.dropIndex ( 
 "cuisine_index" 
 ) 

Delete an index using index definition

db.restaurants.dropIndex ({ 
 "cuisine" 
  
:  
 1 
 }) 

Firebase console

1. In the Firebase console, go to the Firestore Database page.

   Go to Firestore Database

2. Select a database from the list of databases.
3. Click the Indexes tab.
4. In the list of indexes, choose Delete from the More button more_vert for the index you want to delete.
5. Click Delete Index .

gcloud CLI

1. To find the name of the index, use the gcloud firestore indexes composite list command.

   gcloud firestore indexes composite list \
   --database=' DATABASE_ID 
   '

   Replace DATABASE_ID with the database ID.

2. To delete the index, use the gcloud firestore indexes composite delete command.

   gcloud firestore indexes composite delete INDEX_NAME 
   \
   --database=' DATABASE_ID 
   '

   Replace the following:

   • INDEX_NAME : the name of an index
   • DATABASE_ID : a database ID