mongoose 中的 aggregate

Hi,大家好!前陣子在研究專案上從 MySQL migrate 到 MongoDB 的可能性,選擇了 Mongoose 作為這次的主題,這篇文章會帶著大家認識 aggregate 的基本操作。

# 什麼是 Mongoose?

以下為 Mongoose 官方文件上對自己的介紹:

elegant mongodb object modeling for node.js.

writing MongoDB validation, casting and business logic boilerplate is a drag. That's why we wrote Mongoose.

簡單來說,Mongoose 是基於 node.js 的 Object Data Modeling (ODM),讓我們可以優雅的操作 MongoDB 中的資料。

# Aggregate

在進入介紹前,先和大家分享一下在 Mongoose v6.0.7 版本之前使用 Aggregate 會出現下方錯誤資訊。

Error: Method "collection.aggregate()" accepts at most two arguments

MongoInvalidArgumentError: Method "collection.aggregate()" accepts at most two arguments
at Collection.aggregate (/home/dave/projects/asd/node_modules/mongodb/lib/collection.js:367:19)
at NativeCollection.<computed> [as aggregate] (/home/dave/projects/asd/node_modules/mongoose/lib/drivers/node-mongodb-native/collection.js:200:33)
at NativeCollection.Collection.doQueue (/home/dave/projects/asd/node_modules/mongoose/lib/collection.js:135:23)
at /home/dave/projects/asd/node_modules/mongoose/lib/collection.js:82:24
at processTicksAndRejections (internal/process/task_queues.js:77:11)

這個 issue 前幾天被修正了,在實作時可以注意一下 Mongoose 的版本。

# $match & $project

$match 可以讓我們篩選資料,$project 可以組合出最後 output 的資料。
來看看下面的範例。

async function getStoreInfo(storeUuid) {
const ObjectId = mongoose.Types.ObjectId
const [result] = await Store.aggregate([
{ $match: { uuid: ObjectId(storeUuid) } },
{
$project: {
_id: 0,
uuid: "$uuid",
name: "$info.name",
servicePhone: "$info.servicePhone",
address: "$info.address",
},
},
])
return result
}

如果印出上方的 result 會得到下方資料。

{
"uuid": "6123456e1787b96a05123456",
"name": "Error Baker",
"servicePhone": "0287878787",
"address": "台北市中正區羅斯福路一段2號",
}

# $lookup & $unwind

$lookup 可以讓我們使用 JOIN 從其他 collection 拿資料,$unwind 可以拆分資料。
來看看下面的範例。

async function getRoleWithUserByStoreId(storeId) {
const ObjectId = mongoose.Types.ObjectId;
const result = await Role.aggregate([
{ $match: { store: ObjectId(storeId) } },
{
$lookup: {
from: "storeuserroles",
pipeline: [
{
$lookup: {
from: "users", // Collection name
localField: "user",
foreignField: "_id",
as: "user",
},
},
{
$unwind: "$user",
},
{
$project: {
_id: 0,
id: { $toString: "$user._id" },
uuid: { $toString: "$user.uuid" },
name: "$user.name",
email: "$user.email",
phone: "$user.phone",
photoUrl: "$user.photoUrl",
intro: "$user.intro",
},
},
],
as: "users",
},
},
{
$project: {
_id: 0,
id: { $toString: "$_id" },
name: "$name",
users: "$users",
},
},
]);

return result;
}

如果印出上方的 result 會得到下方資料。

[
{
"id": "6123456d87bee84123456887",
"name": "errorBaker筆者",
"users": [
{
"id": "6151234567bee84077654321",
"uuid": "615528412345674072b40887",
"name": "ruofan",
"email": "errorBaker@gmail.com",
"phone": "0987878787",
"photoUrl": null,
"intro": null
},
{
"id": "6123450058bee87072012345",
"uuid": "61552123450ee84072b40870",
"name": "xiang",
"email": "errorBaker@gmail.com",
"phone": "0978787878",
"photoUrl": null,
"intro": null
}
]
},
{
"id": "6123456d87bee80073456801",
"name": "errorBaker管理員",
"users": [
{
"id": "6151234567bee84077654302",
"uuid": "615528412345674070020887",
"name": "huli",
"email": "errorBaker@gmail.com",
"phone": "0987878787",
"photoUrl": null,
"intro": null
},
{
"id": "6123450058bee87072012303",
"uuid": "61552123450ee84000340870",
"name": "clay",
"email": "errorBaker@gmail.com",
"phone": "0978787878",
"photoUrl": null,
"intro": null
}
]
}
]

# $cond

$cond 可以讓我們在 $project 中加上條件判斷。
來看看下面的範例。

async function getUserByType(storeId, userType = null) {
const ObjectId = mongoose.Types.ObjectId;
const result = await StoreUserRole.aggregate([
{ $match: { store: ObjectId(storeId) } },
{
$lookup: {
from: "users",
localField: "user",
foreignField: "_id",
as: "user",
},
},
{ $unwind: "$user" },
{
$lookup: {
from: "roles",
localField: "role",
foreignField: "_id",
as: "role",
},
},
{ $unwind: "$role" },
{
$project: {
_id: 0,
uuid: { $toString: "$user.uuid" },
name: "$user.name",
email: "$user.email",
phone: "$user.phone",
photoUrl: "$user.photoUrl",
intro: "$user.intro",
role: {
id: 0,
id: { $toString: "$role._id" },
name: "$role.name",
isSuperAdmin: {
$cond: [
{ $eq: [userType, null] },
"$role.isSuperAdmin",
"$$REMOVE",
],
},
},
},
},
]);
return result;
}

如果沒有傳入 userType 印出上方的 result 會得到下方資料。

[
{
"role": {
"id": "6878784d58bee12345640801",
"name": "errorBaker筆者",
"isSuperAdmin": false,
},
"uuid": "6878784d58bee12345640887",
"name": "ruofan",
"email": "errorBaker@gmail.com",
"phone": "0987878787",
"photoUrl": null,
"intro": null
},
{
"role": {
"id": "6878784d58bee12345640801",
"name": "errorBaker管理員",
"isSuperAdmin": true,
},
"uuid": "6878784d58bee12345640887",
"name": "huli",
"email": "errorBaker@gmail.com",
"phone": "0988878787",
"photoUrl": null,
"intro": null
}
],

# 小結

Aggregate 中有非常多方法可以讓我們整理與操作資料,在研究的過程中即便踩了一些坑也還是覺得挺有趣的!推薦給大家!

在閱讀文章時如果有遇到什麼問題,或是有什麼建議,都歡迎留言告訴我,謝謝。😃

# 參考資料


關於作者

喜歡有趣的設計

分享文章