Skip to content

Strongly typed ORM for DynamoDB - Built with the single-table-design pattern in mind.

License

Notifications You must be signed in to change notification settings

michael-pont/typedorm

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TypeDORM

Code Style: Google semantic-release

Object Relational mapper for DynamoDB, inspired by typeorm.

TypeDORM is an ORM built from ground up using typescript and latest javascript features to provide an easy gateway when doing complex highly relational data modeling in dynamoDB. TypeDORM is built with single-table-design first in mind, but should work as smoothly with regular entity table <-> design pattern. TypeDORM would have not existed without TypeORM and dynamodb-toolbox, big shout-out to these projects and their awesome contributors.

TypeDORM borrows decorator based syntax from TypeORM and provides fully type safe ORM to with dynamodb. TypeDORM currently only support Data Mapper.

Packages

Package Latest Stable Recent Beta Recent Alpha
@typedorm/common NPM Release NPM Release NPM Release
@typedorm/core NPM Release NPM Release NPM Release
@typedorm/testing NPM Release NPM Release NPM Release
@typedorm/document-client NPM Release NPM Release NPM Release

Branches

Branches Stability
main Stability
beta Stability
alpha Stability

Features

  • AWS SDK V2 and V3 unified support
  • Single-Table design pattern first class support
  • DataMapper development pattern
  • Attribute level per entity transformation, enabled via class-transformer
  • Full type safety
  • Declarative relational schema
  • Entity manager - easy to use findOne, find, count, exists, create, update, and delete operations
  • Transaction manager - easy to use write and read operations
  • Batch manager - powerful write and read operations
  • Scan Manager - powerful find, count, parallelScan and scan operations
  • Safer parallel scan with configurable concurrency control
  • Multiple connections support
  • Code follows all possible best practices when modeling for dynamodb
  • Supports specifying non key attribute as unique, built in pagination support for querying
  • Supports updating Primary key attributes
  • Auto Generated values for attributes
  • Auto update attribute values on UPDATE operations
  • Most type safe, powerful and flexible update expression generator engine
  • Dynamic & Static default values for attributes
  • Complex update, key condition and condition expression all made easy to work with
  • Powerful expression builder to auto generate expressions from input
  • Typescript and javascript support
  • Commonjs and ESM module support out of the box (see more here)

And many more to come.

Getting Started

Installation

  1. Install core and common modules from npm. npm install @typedorm/core @typedorm/common --save

  2. Install AWS SDK for nodejs, TypeDORM uses documentClient to interact with dynamodb. a. When using AWS SDK Version 2 npm install aws-sdk --save

    b. When using AWS SDK Version 3 npm install @aws-sdk/client-dynamodb @aws-sdk/lib-dynamodb --save

  3. Install reflect-metadata shim

    npm install reflect-metadata --save

    and import it as the first thing in node entry file as

    import 'reflect-metadata'

Typescript configuration

If you are using TypeDORM with typescript, make sure you also have below options enabled in tsconfig.json

"emitDecoratorMetadata": true,
"experimentalDecorators": true,

Developing with TypeDORM

Creating Table

First thing to do when working with TypeDORM is to setup dynamodb table config. Currently this needs to be manually setup and have also have it configured in deployed table instance(s).

This guide shows how to setup single-table-design

my-table.ts

import {Table, INDEX_TYPE} from '@typedorm/common';

// create table

const myGlobalTable = new Table({
  name: 'test-table',
  partitionKey: 'PK',
  sortKey: 'SK',
  indexes: {
    GSI1: {
      type: INDEX_TYPE.GSI,
      partitionKey: 'GSI1PK',
      sortKey: 'GSI1SK',
    },
    GSI2: {
      type: INDEX_TYPE.GSI,
      partitionKey: 'GSI2PK',
      sortKey: 'GSI2SK',
    },
    LSI1: {
      type: INDEX_TYPE.LSI,
      sortKey: 'LSI1SK',
    },
  },
});

Note: These indexes must match exactly to what is created in dynamo table instance hosted.

Creating an entity

organisation.entity.ts

import {Attribute, Entity, AutoGenerateAttribute} from '@typedorm/common';
import {AUTO_GENERATE_ATTRIBUTE_STRATEGY} from '@typedorm/common';

@Entity({
  name: 'organisation',
  primaryKey: {
    partitionKey: 'ORG#{{id}}',
    sortKey: 'ORG#{{id}}',
  },
  indexes: {
    // specify GSI1 key - "GSI1" named global secondary index needs to exist in above table declaration
    GSI1: {
      partitionKey: 'ORG#{{id}}#STATUS#{{status}}',
      sortKey: 'ORG#{{id}}#ACTIVE#{{active}}',
      type: INDEX_TYPE.GSI,
    },
    // specify LSI1 key
    LSI1: {
      sortKey: 'TICKETS#UPDATED_AT#{{updatedAt}}',
      type: INDEX_TYPE.LSI,
    },
  },
})
export class Organisation {
  @AutoGenerateAttribute({
    strategy: AUTO_GENERATE_ATTRIBUTE_STRATEGY.UUID4,
  })
  id: string;

  @Attribute()
  name: string;

  @Attribute()
  status: string;

  @Attribute()
  active: boolean;

  @AutoGenerateAttribute({
    strategy: AUTO_GENERATE_ATTRIBUTE_STRATEGY.EPOCH,
    autoUpdate: true, // this will make this attribute and any indexes referencing it auto update for any write operation
  })
  updatedAt: number;
}

Initialize default connection

The connection initialization steps slightly defers when using AWS SDK V2 vs V3.

When using AWS SDK V2
import {createConnection} from '@typedorm/core';
import {DocumentClientV2} from '@typedorm/document-client';
import AWS from 'aws-sdk';

const documentClient = new DocumentClientV2(new AWS.DynamoDB.DocumentClient());

// initialize with specifying list of entities
createConnection({
  table: myGlobalTable,
  entities: [Organisation],
  documentClient, // <-- When documentClient is not provided, TypeDORM defaults to use the DocumentClientV2
});

// or initialize with specifying path match for entities
createConnection({
  table: myGlobalTable,
  entities: 'path-to-entities/*.entity.ts',
  documentClient, // <-- When documentClient is not provided, TypeDORM defaults to use the DocumentClientV2
});
When using AWS SDK V3
import {createConnection} from '@typedorm/core';
import {DocumentClientV3} from '@typedorm/document-client';
import {DynamoDBClient} from '@aws-sdk/client-dynamodb';

const documentClient = new DocumentClientV3(new DynamoDBClient({}));

// initialize with specifying list of entities
createConnection({
  table: myGlobalTable,
  entities: [Organisation],
  documentClient, // <-- When documentClient is not provided, TypeDORM defaults to use the DocumentClientV2
});

// or initialize with specifying path match for entities
createConnection({
  table: myGlobalTable,
  entities: 'path-to-entities/*.entity.ts',
  documentClient, // <-- When documentClient is not provided, TypeDORM defaults to use the DocumentClientV2
});

Working with entity manager

import {getEntityManager} from '@typedorm/core';

const org = new Organisation();
org.name = 'My awesome org';
org.status = 'onboarding';
org.active = true;

const entityManger = getEntityManager();

// create item
const response = await entityManger.create(org);

// get item
const org = await entityManger.findOne(Organisation, {
  id: response.id,
  status: 'onboarding',
  active: true,
});

// delete item
await entityManger.delete(Organisation, {
  id: response.id,
  status: 'onboarding',
  active: true,
});

Table of contents

More

DynamoDB is different, different than most other no-sql databases, and therefore data in dynamodb should be stored the way dynamodb expects to get the most benefits out of it. While doing this development experience suffers and all data can become a huge mess very quickly, this is specially true with single-table-design patten. To resolve this, TypeDORM let's declaratively define schema and later takes control from there to provide best development experience possible.

To find out more about how the data looks like when it is stored in dynamo have a look at this detailed guide.

Current Limitations

  • TypeDORM, at the moment WILL NOT create/update table configuration and must be done separately by the developer.

Sponsors

Nextfaze

Contributions

Please submit an issue for any bugs or ideas here, or you can reach out to me on twitter @whimzy_live.

About

Strongly typed ORM for DynamoDB - Built with the single-table-design pattern in mind.

Resources

License

Security policy

Stars

Watchers

Forks

Packages

No packages published

Languages

  • TypeScript 96.8%
  • JavaScript 1.6%
  • Starlark 1.5%
  • Shell 0.1%