了解 2023 年 Google I/O 大会上介绍的 Firebase 亮点。了解详情

使用聚合查询统计文档数量

聚合查询会处理多个索引条目中的数据,然后返回单个摘要值。

Cloud Firestore 支持 count() 聚合查询。count() 可用于确定集合或查询中的文档数量。服务器会计算数量并且只是将结果(即一个整数)传回您的应用,而不会执行完整的查询,这样做既节省了文档读取费用,又减少了所需传输的字节数。

聚合查询依赖于您的查询已在使用的现有索引配置,并且会根据扫描到的索引条目数量按比例进行扩缩。这意味着,中小规模数据集的聚合过程通常会在 20-40 毫秒内完成,但延迟时间会随着计数项的数量增多而增加。

使用 count() 聚合

请参阅我们在获取数据部分中设置的示例数据。

以下 count() 聚合会返回 cities 集合中的城市总数。

Web 模块化 API

const coll = collection(db, "cities");
const snapshot = await getCountFromServer(coll);
console.log('count: ', snapshot.data().count);
Swift
注意:此产品不适用于 watchOS 和 App Clip 目标。
let query = db.collection("cities")
let countQuery = query.count
do {
    let snapshot = try await countQuery.getAggregation(source: .server)
    print(snapshot.count)
} catch {
    print(error)
}
Objective-C
注意:此产品不适用于 watchOS 和 App Clip 目标。
FIRCollectionReference *query = [self.db collectionWithPath:@"cities"];
[query.count aggregationWithSource:FIRAggregateSourceServer
                        completion:^(FIRAggregateQuerySnapshot *snapshot,
                                     NSError *error) {
    if (error != nil) {
        NSLog(@"Error fetching count: %@", error);
    } else {
        NSLog(@"Cities count: %@", snapshot.count);
    }
}];

Java

Query query = db.collection("cities");
AggregateQuery countQuery = query.count();
countQuery.get(AggregateSource.SERVER).addOnCompleteListener(new OnCompleteListener<AggregateQuerySnapshot>() {
    @Override
    public void onComplete(@NonNull Task<AggregateQuerySnapshot> task) {
        if (task.isSuccessful()) {
            // Count fetched successfully
            AggregateQuerySnapshot snapshot = task.getResult();
            Log.d(TAG, "Count: " + snapshot.getCount());
        } else {
            Log.d(TAG, "Count failed: ", task.getException());
        }
    }
});

Kotlin+KTX

val query = db.collection("cities")
val countQuery = query.count()
countQuery.get(AggregateSource.SERVER).addOnCompleteListener { task ->
    if (task.isSuccessful) {
        // Count fetched successfully
        val snapshot = task.result
        Log.d(TAG, "Count: ${snapshot.count}")
    } else {
        Log.d(TAG, "Count failed: ", task.getException())
    }
}

Dart

// Returns number of documents in users collection
db.collection("users").count().get().then(
      (res) => print(res.count),
      onError: (e) => print("Error completing: $e"),
    );
Go
package firestore

import (
	"context"
	"errors"
	"fmt"
	"io"

	"cloud.google.com/go/firestore"
	firestorepb "cloud.google.com/go/firestore/apiv1/firestorepb"
)

func createCountQuery(w io.Writer, projectID string) error {

	// Instantiate the client
	ctx := context.Background()
	client, err := firestore.NewClient(ctx, projectID)
	if err != nil {
		return err
	}
	defer client.Close()

	collection := client.Collection("users")
	query := collection.Where("born", ">", 1850)

	// `alias` argument--"all"--provides a key for accessing the aggregate query
	// results. The alias value must be unique across all aggregation aliases in
	// an aggregation query and must conform to allowed Document field names.
	//
	// See https://cloud.google.com/firestore/docs/reference/rpc/google.firestore.v1#document for details.
	aggregationQuery := query.NewAggregationQuery().WithCount("all")
	results, err := aggregationQuery.Get(ctx)
	if err != nil {
		return err
	}

	count, ok := results["all"]
	if !ok {
		return errors.New("firestore: couldn't get alias for COUNT from results")
	}

	countValue := count.(*firestorepb.Value)
	fmt.Fprintf(w, "Number of results from query: %d\n", countValue.GetIntegerValue())
	return nil
}
Java
CollectionReference collection = db.collection("cities");
AggregateQuerySnapshot snapshot = collection.count().get().get();
System.out.println("Count: " + snapshot.getCount());
      
Node.js
const collectionRef = db.collection('cities');
const snapshot = await collectionRef.count().get();
console.log(snapshot.data().count);
      
Python
from google.cloud import firestore
from google.cloud.firestore_v1 import aggregation
from google.cloud.firestore_v1.base_query import FieldFilter


def create_count_query(project_id: str) -> None:
    """Builds an aggregate query that returns the number of results in the query.

    Arguments:
      project_id: your Google Cloud Project ID
    """
    client = firestore.Client(project=project_id)

    collection_ref = client.collection("users")
    query = collection_ref.where(filter=FieldFilter("born", ">", 1800))
    aggregate_query = aggregation.AggregationQuery(query)

    # `alias` to provides a key for accessing the aggregate query results
    aggregate_query.count(alias="all")

    results = aggregate_query.get()
    for result in results:
        print(f"Alias of results from query: {result[0].alias}")
        print(f"Number of results from query: {result[0].value}")

count() 聚合会考虑查询的所有过滤条件以及所有 limit 子句。

Web 模块化 API

const coll = collection(db, "cities");
const q = query(coll, where("state", "==", "CA"));
const snapshot = await getCountFromServer(q);
console.log('count: ', snapshot.data().count);
Swift
注意:此产品不适用于 watchOS 和 App Clip 目标。
let query = db.collection("cities").whereField("state", isEqualTo: "CA")
let countQuery = query.count
do {
    let snapshot = try await countQuery.getAggregation(source: .server)
    print(snapshot.count)
} catch {
    print(error)
}
Objective-C
注意:此产品不适用于 watchOS 和 App Clip 目标。
FIRQuery *query =
    [[self.db collectionWithPath:@"cities"]
                 queryWhereField:@"state"
                       isEqualTo:@"CA"];
[query.count aggregationWithSource:FIRAggregateSourceServer
                        completion:^(FIRAggregateQuerySnapshot *snapshot,
                                      NSError *error) {
    if (error != nil) {
        NSLog(@"Error fetching count: %@", error);
    } else {
        NSLog(@"Cities count: %@", snapshot.count);
    }
}];

Java

Query query = db.collection("cities").whereEqualTo("state", "CA");
AggregateQuery countQuery = query.count();
countQuery.get(AggregateSource.SERVER).addOnCompleteListener(new OnCompleteListener<AggregateQuerySnapshot>() {
    @Override
    public void onComplete(@NonNull Task<AggregateQuerySnapshot> task) {
        if (task.isSuccessful()) {
            // Count fetched successfully
            AggregateQuerySnapshot snapshot = task.getResult();
            Log.d(TAG, "Count: " + snapshot.getCount());
        } else {
            Log.d(TAG, "Count failed: ", task.getException());
        }
    }
});

Kotlin+KTX

val query = db.collection("cities").whereEqualTo("state", "CA")
val countQuery = query.count()
countQuery.get(AggregateSource.SERVER).addOnCompleteListener { task ->
    if (task.isSuccessful) {
        // Count fetched successfully
        val snapshot = task.result
        Log.d(TAG, "Count: ${snapshot.count}")
    } else {
        Log.d(TAG, "Count failed: ", task.getException())
    }
}

Dart

// This also works with collectionGroup queries.
db.collection("users").where("age", isGreaterThan: 10).count().get().then(
      (res) => print(res.count),
      onError: (e) => print("Error completing: $e"),
    );
Go
package firestore

import (
	"context"
	"errors"
	"fmt"
	"io"

	"cloud.google.com/go/firestore"
	firestorepb "cloud.google.com/go/firestore/apiv1/firestorepb"
)

func createCountQuery(w io.Writer, projectID string) error {

	// Instantiate the client
	ctx := context.Background()
	client, err := firestore.NewClient(ctx, projectID)
	if err != nil {
		return err
	}
	defer client.Close()

	collection := client.Collection("users")
	query := collection.Where("born", ">", 1850)

	// `alias` argument--"all"--provides a key for accessing the aggregate query
	// results. The alias value must be unique across all aggregation aliases in
	// an aggregation query and must conform to allowed Document field names.
	//
	// See https://cloud.google.com/firestore/docs/reference/rpc/google.firestore.v1#document for details.
	aggregationQuery := query.NewAggregationQuery().WithCount("all")
	results, err := aggregationQuery.Get(ctx)
	if err != nil {
		return err
	}

	count, ok := results["all"]
	if !ok {
		return errors.New("firestore: couldn't get alias for COUNT from results")
	}

	countValue := count.(*firestorepb.Value)
	fmt.Fprintf(w, "Number of results from query: %d\n", countValue.GetIntegerValue())
	return nil
}
Java
CollectionReference collection = db.collection("cities");
Query query = collection.whereEqualTo("state", "CA");
AggregateQuerySnapshot snapshot = query.count().get().get();
System.out.println("Count: " + snapshot.getCount());
      
Node.js
const collectionRef = db.collection('cities');
const query = collectionRef.where('state', '==', 'CA');
const snapshot = await query.count().get();
console.log(snapshot.data().count);
      
Python
from google.cloud import firestore
from google.cloud.firestore_v1 import aggregation
from google.cloud.firestore_v1.base_query import FieldFilter


def create_count_query(project_id: str) -> None:
    """Builds an aggregate query that returns the number of results in the query.

    Arguments:
      project_id: your Google Cloud Project ID
    """
    client = firestore.Client(project=project_id)

    collection_ref = client.collection("users")
    query = collection_ref.where(filter=FieldFilter("born", ">", 1800))
    aggregate_query = aggregation.AggregationQuery(query)

    # `alias` to provides a key for accessing the aggregate query results
    aggregate_query.count(alias="all")

    results = aggregate_query.get()
    for result in results:
        print(f"Alias of results from query: {result[0].alias}")
        print(f"Number of results from query: {result[0].value}")

Cloud Firestore 安全规则针对 count() 聚合查询的工作方式与针对返回文档的常规查询的工作方式相同。换句话说,当且仅当您的规则允许客户端执行某些集合或合集组查询时,客户端也可以对这些查询执行 count() 聚合操作。详细了解 Cloud Firestore 安全规则与查询之间的相互作用

限制

请注意,count() 聚合查询存在以下限制:

  • count() 聚合查询目前只能通过直接服务器响应进行处理。查询只会通过 Cloud Firestore 后端进行处理,并且会跳过本地缓存和任何已缓冲的更新。此行为与 Cloud Firestore 事务内执行的操作相同。您目前无法将 count() 查询与实时监听器和离线查询搭配使用。

  • 如果 count() 聚合无法在 60 秒内完成解析,则会返回 DEADLINE_EXCEEDED 错误。聚合性能取决于您的索引配置以及数据集的规模。

    如果操作无法在 60 秒时限内完成,作为一种可能的解决方法,您可以针对大规模数据集使用计数器

  • count() 聚合会从索引条目中读取数据,并且只对编入索引的字段进行计数。

  • 如果向查询添加 OrderBy 子句,则查询只会对存在排序字段的文档进行计数。

价格

count() 的价格取决于查询匹配的索引条目的数量。您需要为大量匹配的条目支付少量读取费用。

查看更详细的价格信息