🌼 BE TIL Day 20 0408

JB·2022년 4월 11일

Backend Database Encode JWT ManyToMany decode login sign-in

CodeCamp BE 02

목록 보기

19/30

⬇️ Main Note
https://docs.google.com/document/d/1c3LEm0IKmI6y--c3nGdjaxAmZISq7eHd3TfjwD-M1fc/edit

🌿 ManyToMany

Today, I'm trying to make productTags API. This one's in the format of manyToMany in database.
productTags will get the input in the form of array. => productTags: ["#new","#hi"]

tagsRepository should be made and linked with the product by using relations.

⬇️ Source code

async create({ createProductInput }) {
    const {
      productSaleslocation,
      productCategoryId,
      productTags,
      ...restProduct
    } = createProductInput; // dividing createProductInput into several variables so that things can be assigned separately. 

    const result = await this.productSaleslocationRepository.save({
      ...productSaleslocation,
    }); 

    // productTags : ["#wow","#chocolate","#yum"]
    // =-=-=-=-=-= productTags =-=-=-=-=-=
    const result2 = [];
    for (let i = 0; i < productTags.length; i++) {
      const tagName = productTags[i].replace('#', '');

      // ----------- tag overlap check -----------
      const prevTag = await this.productTagRepository.findOne({
        name: tagName,
      }); // Check whether the tag name already exists in database (= original tag)

      if (prevTag) {
        // If original tag exists
        result2.push(prevTag);
      } else {
        // If the original tag doesn't exist
        const newTag = await this.productTagRepository.save({ name: tagName });
        result2.push(newTag);
      }
    }

Memory: Data are saved as variables, so when you shut down the computer, all data are removed.
Disk: Data are saved as text, which is saved inside the computer itself.

☁️ Stateful vs. Stateless

I chose stateless form for now.
-> No need for session savings
The login status is saved in database
-> SessionId = Login Token
*** database is disk-based

🌿 Database Load-balancing

🐚 Vertical Partitioning
Data are saved column by column
ex) Things are saved in big categories: user data, payment data, etc

🐚 Horizontal Partitioning : Sharding
Normalized database(정규화된 DB: 1NF, 2NF...) are saved to reduce the data size.

❗️ Since database is disk-based, the speed of calling the data is slow.

☁️ Redis

So we're gonna use Redis, which is memory-based database.

Basically, Redis is put between backend server and database.
I'm using Redis for calling the user login info : token info.

☁️ JWT - Encryption

Or just save the login info in object-form and save the login info that's encrypted.
Encrypt <==> Decode
ex) ㄱ - r, ㄴ - s, ㄷ - e … // “동동” : “ehdehd”
--> This one's able to be hacked by the hackers.

So JWT is mostly used for encoding the info.

JWT : JSON Web Token --> Decoding website
Here in JWT, key is the string-type password
--> Using this key, only the one who knows the password is able to access for decoding.