如何优化嵌套Firestore查询？不修改数据结构的方案

我需要从Firestore的venues/{id}/offers集合中获取最多10条已激活的文档，按reusable字段倒序排列。同时要针对每条文档，查询users/{uid}/redemptions集合中是否存在对应ID的文档。当前查询总耗时3-4秒，时间统计显示两次查询各耗时1300-1500毫秒，希望在不修改数据结构的前提下优化性能。

原代码 #

export const getOffers = async (id, uid) => {
  const ref = collection(FIREBASE_FIRESTORE, "venues", id, "offers");
  const q = query(
    ref,
    where("active", "==", true),
    orderBy("reusable", "desc"),
    limit(10)
  );
  const snapshot = await getDocs(q);
  const promises = snapshot.docs.map(async (document) => {
    const ref = doc(
      FIREBASE_FIRESTORE,
      "users",
      uid,
      "redemptions",
      document.id
    );
    const redemptionSnapshot = await getDoc(ref);
    return {
      id: document.id,
      used: redemptionSnapshot.exists(),
      ...document.data(),
    };
  });
  const data = await Promise.all(promises);
  return data;
};

添加时间统计后的代码及结果 #

export const getOffers = async (id, uid) => {
  const startFirst = performance.now();
  const ref = collection(FIREBASE_FIRESTORE, "venues", id, "offers");
  const q = query(
    ref,
    where("active", "==", true),
    orderBy("reusable", "desc"),
    limit(10)
  );
  const snapshot = await getDocs(q);
  const endFirst = performance.now();
  console.log("FIRST CALL TAKES: " + (endFirst - startFirst) + " MILLISECONDS");
  const startSecond = performance.now();
  const promises = snapshot.docs.map(async (document) => {
    const ref = doc(
      FIREBASE_FIRESTORE,
      "users",
      uid,
      "redemptions",
      document.id
    );
    const redemptionSnapshot = await getDoc(ref);
    const endSecond = performance.now();
    console.log(
      "SECOND CALL TAKES: " + (endSecond - startSecond) + " MILLISECONDS"
    );
    return {
      id: document.id,
      used: redemptionSnapshot.exists(),
      ...document.data(),
    };
  });
  const data = await Promise.all(promises);
  return data;
};

输出结果：

FIRST CALL TAKES: 1306.4292080402374 MILLISECONDS
SECOND CALL TAKES: 1313.7304170131683 MILLISECONDS

所有查询耗时均在1300-1500毫秒左右。

优化方案（不修改数据结构） #

1. 批量查询用户兑换记录 #

将循环调用的getDoc改为批量查询，把10次独立请求合并为1次，减少网络往返开销：

import { documentId } from "firebase/firestore";

export const getOffers = async (id, uid) => {
  // 获取优惠活动列表
  const offersRef = collection(FIREBASE_FIRESTORE, "venues", id, "offers");
  const q = query(
    offersRef,
    where("active", "==", true),
    orderBy("reusable", "desc"),
    limit(10)
  );
  const offersSnapshot = await getDocs(q);
  const offerIds = offersSnapshot.docs.map(doc => doc.id);
  
  // 批量查询对应的兑换记录
  const redemptionsRef = collection(FIREBASE_FIRESTORE, "users", uid, "redemptions");
  const redemptionQ = query(redemptionsRef, where(documentId(), "in", offerIds));
  const redemptionSnapshot = await getDocs(redemptionQ);
  
  // 构建已使用的优惠ID集合
  const usedOfferIds = new Set(redemptionSnapshot.docs.map(doc => doc.id));
  
  // 组装最终数据
  return offersSnapshot.docs.map(doc => ({
    id: doc.id,
    used: usedOfferIds.has(doc.id),
    ...doc.data()
  }));
};

2. 启用本地持久化缓存 #

如果是客户端环境（Web/移动端），开启Firestore持久化缓存，重复查询可直接从本地读取：

// 初始化Firestore时配置持久化
await initializeFirestore(app, {
  persistence: true
});

3. 确认复合索引存在 #

确保venues/{id}/offers集合有active（过滤）+reusable（排序）的复合索引。Firebase控制台会在查询报错时提示缺失索引，直接按提示创建即可，避免查询时的索引临时构建开销。

4. 减少不必要的数据传输 #

如果offers文档包含大量非必要字段，使用select方法仅获取需要的字段，降低数据传输体积：

const q = query(
  offersRef,
  where("active", "==", true),
  orderBy("reusable", "desc"),
  limit(10),
  select("reusable", "title", "description") // 只获取业务需要的字段
);

内容的提问来源于stack exchange，提问作者David Henry