Skip to main content

How to model NoSQL one-to-one relationships

· 8 min read
Garren

Tigris provides a transactional NoSQL document database, this allows you to create NoSQL data models representing relations in a few different ways to get the best performance, consistency and ease of use in your application.

In this blog, I show you how to model NoSQL one-to-one relationships using the embedded document and sub-document NoSQL design pattern. We'll get hands-on with Tigris to achieve this, however the NoSQL data modeling design patterns are applicable to other NoSQL databases. We will combine this with Tigris's transactions to make sure that all updates are consistent. I will show how to model NoSQL one-to-many and NoSQL many-to-many relationships in future posts as part of the NoSQL Data Modeling Series.

NoSQL one-to-one data modeling

Getting started​

To follow along, you will need to get set up with Tigris. This is a quick and easy process. The Tigris Docs have a Quickstart to show you how to sign up. Once you have signed up, create a new project called relations. Once created, run the following command to create a local application that we can use throughout this guide:

$ npx create-tigris-app@latest --project relations --example playground

Follow the prompts to add your clientId and clientSecret.

Embedded one-to-one NoSQL data modeling​

The first way to model NoSQL one-to-one relationships between models is to embed the model in the document. Let's do that in Tigris, by modeling a User and a Location. In our project, there is already a user.ts file under src/db/models. Change that file to contain two classes like this:

import {
  Field,
  PrimaryKey,
  TigrisCollection,
  TigrisDataTypes,
} from "@tigrisdata/core";

export class Location {
  @Field()
  country: string;

  @Field()
  address: string;

  @Field()
  code: string;
}

@TigrisCollection("user")
export class User {
  @PrimaryKey({ order: 1, autoGenerate: false })
  email: string;

  @Field()
  name: string;

  @Field({ elements: Location })
  location: Location;
}

This creates a Location object embedded inside the User schema. This allows us to store and read the Location and User in one query. The User's email address is the primary key.

Let's create a few users and query them. In the src/index.ts file, let's change the file to look like this:

import { Tigris } from "@tigrisdata/core";
import { User } from "./db/models/user";

// setup client
const tigrisClient = new Tigris();
async function setup() {
  // ensure branch exists, create it if it needs to be created dynamically
  await tigrisClient.getDatabase().initializeBranch();
  // register schemas
  await tigrisClient.registerSchemas([User]);
}

async function main() {
  await setup();
  const db = await tigrisClient.getDatabase();
  const userCollection = await db.getCollection<User>(User);

  await userCollection.insertMany([
    {
      email: "eddie@temp.com",
      name: "Eddie Vedder",
      location: {
        country: "USA",
        address: "1 Red Road",
        code: "8878",
      },
    },
    {
      email: "Jimi@temp.com",
      name: "Jimi Hendrix",
      location: {
        country: "USA",
        address: "13 Yellow Road",
        code: "1234",
      },
    },
    {
      email: "johnny@temp.com",
      name: "Johnny Clegg",
      location: {
        country: "RSA",
        address: "87 Green Road",
        code: "9087",
      },
    },
  ]);

  let user = await userCollection.findOne({ filter: { name: "Eddie Vedder" } });
  console.table(user);

  let userCursor = await userCollection.findMany({
    filter: { "location.country": "USA" },
  });

  for await (user of userCursor) {
    console.table(user);
  }
}

main()
  .then(async () => {
    console.log("Query complete ...");
    process.exit(0);
  })
  .catch(async (e) => {
    console.error(e);
    process.exit(1);
  });

Let's unpack what is happening. In our setup function we are setting up the Tigris NoSQL database, creating our user collection and defining the schema for that collection.

Then in our main function, we get the database object and then fetch our user collection object so that we can work with our collection. We create 3 new users. And then show how to query for one on line 48, or if we want to query many of them based on a filter, we do that on line 51.

tip

In the above example, we call the setup() function to check the database and update the schema. In a production application this only needs to be called when the schema has changed.

Reference one-to-one NoSQL relationships​

There can be cases where our embedded objects are very large and we do not need to load them every time. In these cases, we can split the document into its own collection and fetch it when we need it. Lets add a short biography about the users we have added above and put it in a separate collection so that we can load it when we need to.

We are going to add a new biography object in our user.ts so the file now looks like this:

import {
  Field,
  PrimaryKey,
  TigrisCollection,
  TigrisDataTypes,
} from "@tigrisdata/core";

export class Location {
  @Field()
  country: string;

  @Field()
  address: string;

  @Field()
  code: string;
}

@TigrisCollection("biography")
export class Biography {
  @PrimaryKey(TigrisDataTypes.UUID, { order: 1, autoGenerate: true })
  id?: string;

  @Field()
  userId: string;

  @Field()
  description: string;
}

@TigrisCollection("user")
export class User {
  @PrimaryKey({ order: 1, autoGenerate: false })
  email: string;

  @Field()
  name: string;

  @Field({ elements: Location })
  location: Location;
}

As you can see, we have defined a new class, Biography and given it the collection name biography. It also has a field userId, which is a reference back to the User and will contain the user's email address. We can now update our creating and reading of users to include the biography information.

Change the index.tx to look like this:

import { Tigris } from "@tigrisdata/core";
import { User, Biography } from "./db/models/user";

// setup client
const tigrisClient = new Tigris();
async function setup() {
  // ensure branch exists, create it if it needs to be created dynamically
  await tigrisClient.getDatabase().initializeBranch();
  // register schemas
  await tigrisClient.registerSchemas([User, Biography]);
}

async function main() {
  await setup();
  const db = await tigrisClient.getDatabase();
  const userCollection = await db.getCollection<User>(User);
  const bioCollection = await db.getCollection<Biography>(Biography);

  const tx = await db.beginTransaction();
  await userCollection.insertMany(
    [
      {
        email: "eddie@temp.com",
        name: "Eddie Vedder",
        location: {
          country: "USA",
          address: "1 Red Road",
          code: "8878",
        },
      },
      {
        email: "Jimi@temp.com",
        name: "Jimi Hendrix",
        location: {
          country: "USA",
          address: "13 Yellow Road",
          code: "1234",
        },
      },
      {
        email: "johnny@temp.com",
        name: "Johnny Clegg",
        location: {
          country: "RSA",
          address: "87 Green Road",
          code: "9087",
        },
      },
    ],
    tx
  );

  await bioCollection.insertMany(
    [
      {
        userId: "eddie@temp.com",
        description:
          "Eddie Jerome Vedder is an American singer, musician, and songwriter best known as the lead vocalist and one of three guitarists of the rock band Pearl Jam.",
      },
      {
        userId: "jimi@temp.com",
        description:
          'James Marshall "Jimi" Hendrix was an American guitarist, singer and songwriter.',
      },
      {
        userId: "johnny@temp.com",
        description:
          "Jonathan Paul Clegg, OBE OIS was a South African musician, singer-songwriter, dancer, anthropologist and anti-apartheid activist, some of whose work was in musicology focused on the music of indigenous South African peoples.",
      },
    ],
    tx
  );
  await tx.commit();

  let user = await userCollection.findOne({ filter: { name: "Eddie Vedder" } });
  console.table(user);

  const readTx = await db.beginTransaction();
  let userCursor = await userCollection.findMany(
    {
      filter: { "location.country": "USA" },
    },
    readTx
  );

  for await (user of userCursor) {
    let userBio = await bioCollection.findOne(
      { filter: { userId: user.email } },
      readTx
    );
    console.log(user);
    console.log(userBio);
  }
}

main()
  .then(async () => {
    console.log("main complete ...");
    process.exit(0);
  })
  .catch(async (e) => {
    console.error(e);
    process.exit(1);
  });

We have added the biography as a collection in the setup function. Then in main, we created a transaction before we insert our user and biography documents. We pass this transaction to both insert functions so that both insert operations will either complete successfully, or the transaction will roll the inserts back. This makes sure that writing to both collections is safe and consistent. We never have to worry about our data not being accurate or correct.

Now we can read from the two collections. We do this at Line 33. We start another read transaction and first read from the userCollection as we stream those results back we also fetch the biography of the user from the bioCollection. Again we use a transaction to make sure that we are reading a consistent snapshot of the data and that no modifications will affect the queries.

This blog post is a fast overview of NoSQL data modeling approaches to one-to-one relations. We also demonstrated using the power of Tigris's transactions to give you a few options depending on your use case. In the next blog post, I will show how to model NoSQL one-to-many relations.

Stay connected

Make sure you don't miss the next post in the series by subscribing to the Tigris Newsletter: