Mongoose 使用 getter 函数填充虚拟字段不起作用答案

【问题标题】：Mongoose populate virtual field with a getter function is not workingMongoose 使用 getter 函数填充虚拟字段不起作用
【发布时间】：2019-09-22 18:37:07
【问题描述】：

我正在使用 getter 函数填充 mongoose 模式中的虚拟字段。它应该在没有任何错误的情况下填充该字段，但它会抛出错误

MongooseError: 如果要填充虚拟，则必须设置 localField 和 foreignField 选项

当我们使用 getter 填充字段时，不需要 localfield 和 foriegn 字段

const DashboardSchema = new Schema({
    userId: {
        type: SchemaTypes.ObjectId
    }
}, { toJSON: { virtuals: true } });

DashboardSchema.virtual('TopReports').get(function () {
   return TopReports.find({ userId: this.userId }).sort('-date').limit(10);
})

【问题讨论】：

标签： mongodb mongoose

【解决方案1】：

如果您想要一个带有sort 和limit 的“虚拟填充”，那么您实际上并不是这样做的。您创建的只是一个“虚拟吸气剂”，它实际上是返回异步函数的结果。您可以使用它，但管理解析返回的Promise 会很棘手，而且它与populate() 无关，这是您引发错误的地方。

也有不同的选择。

Mongoose 虚拟填充

对于这个最接近你尝试的东西，你想要这样的东西：

const dashboardSchema = new Schema({
  userId: Schema.Types.ObjectId
},
{
  toJSON: { virtuals: true },
  toObject: { virtuals: true }
});

dashboardSchema.virtual('TopReports', {
  ref: 'Report',
  localField: 'userId',
  foreignField: 'userId',
  options: { sort: { date: -1 }, limit: 10 } // optional - could add with populate
});

const reportSchema = new Schema({
  userId: Schema.Types.ObjectId,
  seq: Number,
  date: Date
});

const Dashboard = mongoose.model('Dashboard', dashboardSchema);
const Report = mongoose.model('Report', reportSchema);

这实际上与"Populate Virtuals" 上的文档示例几乎相同，因为那里的给定示例还包括options，这与将选项传递给填充方法本身相同。当您在 virtual 中设置 options 时，您只需要像这样调用：

let result = await Dashboard.findOne().populate('TopReports');

sort 和 limit 的默认设置会在执行 populate() 时自动应用于此“虚拟”字段。如果您选择不包含 options，您只需手动添加选项：

let result2 = await Dashboard.findOne().populate({
  path: 'TopReports',
  options: { sort: '-date', limit: 5 }
});
log(result2);

重要 - 在virtual 中设置options 将始终覆盖任何options 传递给populate()，如上所示。如果你想在不同的请求上使用不同的options，那么你调用上面的方法而不是定义附加到架构的virtual方法。

这就是你真正需要做的。当然，定义包括localField 和foreignField 以及ref，因此populate() 调用知道从哪里获取数据以及与哪些字段相关。还有一个可选 justOne 用于区分单数和Array 结果，以及一些其他选项。

MongoDB $查找

这里的另一个选项是 MongoDB 基本上内置了相同的功能，除了这是 单个请求 而不是 populate() 实际上是 多个请求 为了从不同的集合返回数据：

   let result = await Dashboard.aggregate([
      { "$lookup": {
        "from": Report.collection.name,
        "let": { "userId": "$userId" },
        "pipeline": [
          { "$match": { "$expr": { "$eq": [ "$userId", "$$userId" ] } } },
          { "$sort": { "date": -1 } },
          { "$limit": 10 }
        ],
        "as": "TopReports"
      }}
    ]);

与populate() 实际发出的“许多”find() 请求相比，这是一个具有单一响应的请求。结果相同，只是使用了$lookup 的“子管道”形式，以便将$sort 和$limit 应用于返回的相关项数组。

通过一些工作，您甚至可以从 mongoose 模式中检查模式定义（包括定义的 virtual ）并构造相同的 $lookup 语句。在Querying after populate in Mongoose 上有这个“模式检查”的基本演示。

所以这取决于最适合您的需求。我建议同时尝试，甚至对应用程序性能进行基准测试。

作为一个完整的演示，这里是一个示例清单。它插入了 20 个东西，只返回数组中最近的 10 个结果：

const { Schema, Types: { ObjectId } } = mongoose = require('mongoose');

const uri = 'mongodb://localhost:27017/test';
const opts = { useNewUrlParser: true };

mongoose.set('useFindAndModify', false);
mongoose.set('useCreateIndex', true);
mongoose.set('debug', true);

const dashboardSchema = new Schema({
  userId: Schema.Types.ObjectId
},
{
  toJSON: { virtuals: true },
  toObject: { virtuals: true }
});

dashboardSchema.virtual('TopReports', {
  ref: 'Report',
  localField: 'userId',
  foreignField: 'userId',
  options: { sort: { date: -1 }, limit: 10 } // optional - could add with populate
});

const reportSchema = new Schema({
  userId: Schema.Types.ObjectId,
  seq: Number,
  date: Date
});

const Dashboard = mongoose.model('Dashboard', dashboardSchema);
const Report = mongoose.model('Report', reportSchema);

const log = data => console.log(JSON.stringify(data, undefined, 2));

(async function() {

  try {

    const conn = await mongoose.connect(uri, opts);

    await Promise.all(
      Object.entries(conn.models).map(([k,m]) => m.deleteMany())
    );

    // Insert some things
    let { userId } = await Dashboard.create({ userId: new ObjectId() });

    const oneDay = ( 1000 * 60 * 60 * 24 );
    const baseDate = Date.now() - (oneDay * 30); // 30 days ago

    await Report.insertMany(
      [ ...Array(20)]
        .map((e,i) => ({ userId, seq: i+1, date: baseDate + ( oneDay * i ) }))
    );

    // Virtual populate
    let popresult = await Dashboard.findOne().populate('TopReports');
    log(popresult);


    // Aggregate $lookup
    let result = await Dashboard.aggregate([
      { "$lookup": {
        "from": Report.collection.name,
        "let": { "userId": "$userId" },
        "pipeline": [
          { "$match": { "$expr": { "$eq": [ "$userId", "$$userId" ] } } },
          { "$sort": { "date": -1 } },
          { "$limit": 10 }
        ],
        "as": "TopReports"
      }}
    ]);

    log(result);


  } catch (e) {
    console.error(e)
  } finally {
    mongoose.disconnect()
  }

})()

还有输出：

Mongoose: dashboards.deleteMany({}, {})
Mongoose: reports.deleteMany({}, {})
Mongoose: dashboards.insertOne({ _id: ObjectId("5cce3d9e16302f32acb5c572"), userId: ObjectId("5cce3d9e16302f32acb5c571"), __v: 0 })
Mongoose: reports.insertMany([ { _id: 5cce3d9e16302f32acb5c573, userId: 5cce3d9e16302f32acb5c571, seq: 1, date: 2019-04-05T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c574, userId: 5cce3d9e16302f32acb5c571, seq: 2, date: 2019-04-06T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c575, userId: 5cce3d9e16302f32acb5c571, seq: 3, date: 2019-04-07T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c576, userId: 5cce3d9e16302f32acb5c571, seq: 4, date: 2019-04-08T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c577, userId: 5cce3d9e16302f32acb5c571, seq: 5, date: 2019-04-09T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c578, userId: 5cce3d9e16302f32acb5c571, seq: 6, date: 2019-04-10T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c579, userId: 5cce3d9e16302f32acb5c571, seq: 7, date: 2019-04-11T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c57a, userId: 5cce3d9e16302f32acb5c571, seq: 8, date: 2019-04-12T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c57b, userId: 5cce3d9e16302f32acb5c571, seq: 9, date: 2019-04-13T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c57c, userId: 5cce3d9e16302f32acb5c571, seq: 10, date: 2019-04-14T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c57d, userId: 5cce3d9e16302f32acb5c571, seq: 11, date: 2019-04-15T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c57e, userId: 5cce3d9e16302f32acb5c571, seq: 12, date: 2019-04-16T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c57f, userId: 5cce3d9e16302f32acb5c571, seq: 13, date: 2019-04-17T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c580, userId: 5cce3d9e16302f32acb5c571, seq: 14, date: 2019-04-18T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c581, userId: 5cce3d9e16302f32acb5c571, seq: 15, date: 2019-04-19T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c582, userId: 5cce3d9e16302f32acb5c571, seq: 16, date: 2019-04-20T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c583, userId: 5cce3d9e16302f32acb5c571, seq: 17, date: 2019-04-21T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c584, userId: 5cce3d9e16302f32acb5c571, seq: 18, date: 2019-04-22T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c585, userId: 5cce3d9e16302f32acb5c571, seq: 19, date: 2019-04-23T01:34:22.554Z, __v: 0 }, { _id: 5cce3d9e16302f32acb5c586, userId: 5cce3d9e16302f32acb5c571, seq: 20, date: 2019-04-24T01:34:22.554Z, __v: 0 } ], {})
Mongoose: dashboards.findOne({}, { projection: {} })
Mongoose: reports.find({ userId: { '$in': [ ObjectId("5cce3d9e16302f32acb5c571") ] } }, { sort: { date: -1 }, limit: 10, projection: {} })
{
  "_id": "5cce3d9e16302f32acb5c572",
  "userId": "5cce3d9e16302f32acb5c571",
  "__v": 0,
  "TopReports": [
    {
      "_id": "5cce3d9e16302f32acb5c586",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 20,
      "date": "2019-04-24T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c585",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 19,
      "date": "2019-04-23T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c584",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 18,
      "date": "2019-04-22T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c583",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 17,
      "date": "2019-04-21T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c582",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 16,
      "date": "2019-04-20T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c581",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 15,
      "date": "2019-04-19T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c580",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 14,
      "date": "2019-04-18T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c57f",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 13,
      "date": "2019-04-17T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c57e",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 12,
      "date": "2019-04-16T01:34:22.554Z",
      "__v": 0
    },
    {
      "_id": "5cce3d9e16302f32acb5c57d",
      "userId": "5cce3d9e16302f32acb5c571",
      "seq": 11,
      "date": "2019-04-15T01:34:22.554Z",
      "__v": 0
    }
  ],
  "id": "5cce3d9e16302f32acb5c572"
}
Mongoose: dashboards.aggregate([ { '$lookup': { from: 'reports', let: { userId: '$userId' }, pipeline: [ { '$match': { '$expr': { '$eq': [ '$userId', '$$userId' ] } } }, { '$sort': { date: -1 } }, { '$limit': 10 } ], as: 'TopReports' } } ], {})
[
  {
    "_id": "5cce3d9e16302f32acb5c572",
    "userId": "5cce3d9e16302f32acb5c571",
    "__v": 0,
    "TopReports": [
      {
        "_id": "5cce3d9e16302f32acb5c586",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 20,
        "date": "2019-04-24T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c585",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 19,
        "date": "2019-04-23T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c584",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 18,
        "date": "2019-04-22T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c583",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 17,
        "date": "2019-04-21T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c582",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 16,
        "date": "2019-04-20T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c581",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 15,
        "date": "2019-04-19T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c580",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 14,
        "date": "2019-04-18T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c57f",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 13,
        "date": "2019-04-17T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c57e",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 12,
        "date": "2019-04-16T01:34:22.554Z",
        "__v": 0
      },
      {
        "_id": "5cce3d9e16302f32acb5c57d",
        "userId": "5cce3d9e16302f32acb5c571",
        "seq": 11,
        "date": "2019-04-15T01:34:22.554Z",
        "__v": 0
      }
    ]
  }
]

错误的方式

只是为了演示“getter”方法有什么问题，下面是一个示例清单，显示了实际解决每个返回对象上返回的Promise：

const { Schema, Types: { ObjectId } } = mongoose = require('mongoose');

const uri = 'mongodb://localhost:27017/test';
const opts = { useNewUrlParser: true };

mongoose.set('useFindAndModify', false);
mongoose.set('useCreateIndex', true);
mongoose.set('debug', true);

const dashboardSchema = new Schema({
  userId: Schema.Types.ObjectId
},
{
  toJSON: { virtuals: true },
  toObject: { virtuals: true }
});

dashboardSchema.virtual('TopReports').get(function() {
  return Report.find({ userId: this.userId }).sort("-date").limit(10);
});

const reportSchema = new Schema({
  userId: Schema.Types.ObjectId,
  seq: Number,
  date: Date
});

const Dashboard = mongoose.model('Dashboard', dashboardSchema);
const Report = mongoose.model('Report', reportSchema);

const log = data => console.log(JSON.stringify(data, undefined, 2));

(async function() {

  try {

    const conn = await mongoose.connect(uri, opts);

    await Promise.all(
      Object.entries(conn.models).map(([k,m]) => m.deleteMany())
    );

    // Insert some things
    let { userId } = await Dashboard.create({ userId: new ObjectId() });

    const oneDay = ( 1000 * 60 * 60 * 24 );
    const baseDate = Date.now() - (oneDay * 30); // 30 days ago

    await Report.insertMany(
      [ ...Array(20)]
        .map((e,i) => ({ userId, seq: i+1, date: baseDate + ( oneDay * i ) }))
    );

    // Mimic the virtual populate with the getter
    let results = await Dashboard.find();
    for ( let r of results ) {
      let obj = { ...r.toObject() };        // copy the plain object data only
      obj.TopReports = await r.TopReports;  // Resolve the Promise
      log(obj);
    }

  } catch (e) {
    console.error(e)
  } finally {
    mongoose.disconnect()
  }

})()

然后输出：

Mongoose: dashboards.deleteMany({}, {})
Mongoose: reports.deleteMany({}, {})
Mongoose: dashboards.insertOne({ _id: ObjectId("5cce45193134aa37e88c4114"), userId: ObjectId("5cce45193134aa37e88c4113"), __v: 0 })
Mongoose: reports.insertMany([ { _id: 5cce45193134aa37e88c4115, userId: 5cce45193134aa37e88c4113, seq: 1, date: 2019-04-05T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4116, userId: 5cce45193134aa37e88c4113, seq: 2, date: 2019-04-06T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4117, userId: 5cce45193134aa37e88c4113, seq: 3, date: 2019-04-07T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4118, userId: 5cce45193134aa37e88c4113, seq: 4, date: 2019-04-08T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4119, userId: 5cce45193134aa37e88c4113, seq: 5, date: 2019-04-09T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c411a, userId: 5cce45193134aa37e88c4113, seq: 6, date: 2019-04-10T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c411b, userId: 5cce45193134aa37e88c4113, seq: 7, date: 2019-04-11T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c411c, userId: 5cce45193134aa37e88c4113, seq: 8, date: 2019-04-12T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c411d, userId: 5cce45193134aa37e88c4113, seq: 9, date: 2019-04-13T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c411e, userId: 5cce45193134aa37e88c4113, seq: 10, date: 2019-04-14T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c411f, userId: 5cce45193134aa37e88c4113, seq: 11, date: 2019-04-15T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4120, userId: 5cce45193134aa37e88c4113, seq: 12, date: 2019-04-16T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4121, userId: 5cce45193134aa37e88c4113, seq: 13, date: 2019-04-17T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4122, userId: 5cce45193134aa37e88c4113, seq: 14, date: 2019-04-18T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4123, userId: 5cce45193134aa37e88c4113, seq: 15, date: 2019-04-19T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4124, userId: 5cce45193134aa37e88c4113, seq: 16, date: 2019-04-20T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4125, userId: 5cce45193134aa37e88c4113, seq: 17, date: 2019-04-21T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4126, userId: 5cce45193134aa37e88c4113, seq: 18, date: 2019-04-22T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4127, userId: 5cce45193134aa37e88c4113, seq: 19, date: 2019-04-23T02:06:17.518Z, __v: 0 }, { _id: 5cce45193134aa37e88c4128, userId: 5cce45193134aa37e88c4113, seq: 20, date: 2019-04-24T02:06:17.518Z, __v: 0 } ], {})
Mongoose: dashboards.find({}, { projection: {} })
Mongoose: reports.find({ userId: ObjectId("5cce45193134aa37e88c4113") }, { sort: { date: -1 }, limit: 10, projection: {} })
{
  "_id": "5cce45193134aa37e88c4114",
  "userId": "5cce45193134aa37e88c4113",
  "__v": 0,
  "TopReports": [
    {
      "_id": "5cce45193134aa37e88c4128",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 20,
      "date": "2019-04-24T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c4127",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 19,
      "date": "2019-04-23T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c4126",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 18,
      "date": "2019-04-22T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c4125",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 17,
      "date": "2019-04-21T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c4124",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 16,
      "date": "2019-04-20T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c4123",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 15,
      "date": "2019-04-19T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c4122",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 14,
      "date": "2019-04-18T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c4121",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 13,
      "date": "2019-04-17T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c4120",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 12,
      "date": "2019-04-16T02:06:17.518Z",
      "__v": 0
    },
    {
      "_id": "5cce45193134aa37e88c411f",
      "userId": "5cce45193134aa37e88c4113",
      "seq": 11,
      "date": "2019-04-15T02:06:17.518Z",
      "__v": 0
    }
  ],
  "id": "5cce45193134aa37e88c4114"
}

当然需要为每个返回的文档解析getter返回的Promise。相比之下，populate() 使用$in 以通过单个请求返回find() 中所有结果的匹配条目，这将为每个Dashboard 文档而不是单个@ 发出一个新的find() 987654372@ 基于结果中每个 Dashboard 文档中找到的所有 userId 值。

基本上，与populate() 或$lookup 不同，您实际上是将“加入”逻辑拆分为控制流的部分，它确实不属于这些部分，并且变得难以管理以及生成更多请求返回服务器。

【讨论】：