CODEWITHSUNDEEP
mongodbmongoosequery-optimization
Fuangel Sikarin
Fuangel Sikarin

Reputation: 77

MongoDB slow to query 1M data

1st of all sorry for my english.

For now my project have 1 million chat record. So I want to query Chats collection from mongodb but It's take time to query and some time take 100% of cpu. So I want to know how to optimize query.

After I research maybe It should use index and lean() but I'm not clear how to using this with my data.

Here my Chat data look like

{
    "_id" : ObjectId("602c53cad1dc22359729bb62"),
    "textMessage" : "1",
    "imageNames" : [],
    "uuid" : "a9y10mfqtokl8mtk0l",
    "sendBy" : "a9y10m9s2skkwu564g",
    "sendTo" : "a9y10m9s2skkwvo7xl",
    "createDateTimestamp" : 1613517770901.0,
    "__v" : 0
}

Here my query function

  Promise.all([
    UserModel.findOne({ uuid: selfUuid }).lean(),
    UserModel.findOne({ uuid: chatWithUuid }).lean(),
    Chat.find({
      $or: [
        { sendBy: uuid, sendTo: chatWithUuid },
        { sendBy: chatWithUuid, sendTo: uuid },
      ],
    })
      .lean()
      .sort({ createDateTimestamp: -1 })
  ]).then((result) => {
    const selfUser = result[0]
    const chatWithUser = result[1]
    const chatMessages = result[2]
    const responseData = chatMessages.map(({ sendBy, sendTo, ...props }) => {
      if (sendTo === selfUser.uuid) {
        return {
          ...props,
          sendTo: selfUser,
          sendBy: chatWithUser,
        }
      } else {
        return {
          ...props,
          sendTo: chatWithUser,
          sendBy: selfUser,
        }
      }
    })
    response.success(res, responseData)
  }).catch((e) => {
    response.fail(res, e, 'fail to getSingleChat')
  })

Upvotes: 0

Views: 260

Answers (2)

Burak Deniz
Burak Deniz

Reputation: 161

If you give index to collection for createDateTimestamp in mongodb. You don't need .sort({ createDateTimestamp: -1 }) anymore.

Upvotes: 1

R2D2
R2D2

Reputation: 10737

Based on your query function and model you can create following two indexes to speed up your searches:

1) {uuid:1}
2) {sendBy:1,sendTo:1}

Upvotes: 1

Related Questions