Navigation
Overview
Installation
Setup
Usage
Upcoming Features
License
Overview
This package is a simple wrapper for common functionality you want when using Redshift. It can do
- Redshift connections & querying
- Creating and running migrations
- Create and manage models
- CRUD API with ORM wrapper with type validation
Warning!!!!!! This is new and still under development. The API is bound to change. Use at your own risk.
Installation
Install the package by running
npm install node-redshift
Link to npm repository https://www.npmjs.com/package/node-redshift
Setup
The code to connect to redshift should be something like this:
//redshift.jsvar Redshift = ; var client = user: user database: database password: password port: port host: host; // The values passed in to the options object will be the difference between a connection pool and raw connectionvar redshiftClient = client options; moduleexports = redshiftClient;
There are two ways to setup a connection to redshift.
- Connection Pooling - you can open a connection pool and open connections to Redshift which will be managed by pg-pool (https://github.com/brianc/node-pg-pool)
- Raw Connection - a one time connection you must manually initialize and close to run queries
***By default node-redshift uses connection pooling
Raw Connection
Pass in the rawConnection parameter in the redshift instantiation options to specify a raw connection. Raw connections need extra code to specify when to connect and disconnect from Redshift. Here's an example of the raw connection query
var redshiftClient = client rawConnection: true;
Connection Pooling
Connection pooling works by default with no extra configuration. Here's an example of connection pooling
Setup Options
There are two options that can be passed into the options object in the Redshift constructor.
Option | Type | Description |
---|---|---|
rawConnection | Boolean | If you want a raw connection, pass true with this option |
longStackTraces | Boolean | Default: true. If you want to disable bluebird's longStackTraces, pass in false |
Usage
Query API
CLI
Models
ORM
Query API
Please see examples/ folder for full code examples using both raw connections and connection pools.
For those looking for a library to build robust, injection safe SQL, I like sql-bricks to build query strings.
Both Raw Connections and Connection Pool connections have two query functions that are bound to the initialized Redshift object: query()
and a parameterizedQuery()
.
All query()
and parameterizedQuery()
functions support both callback and promise style. If there's a function as a third argument, the callback will fire. If there's no third function argument, but instead (query, [options]).then({})... the promise will fire.
//raw connectionvar redshiftClient = ; redshiftClient;
//connection poolvar redshiftClient = ; // options is an optional object with one property so far {raw: true} returns // just the data from redshift. {raw: false} returns the data with the pg objectredshiftClient;//instead of promises you can also use callbacks to get the data
Parameterized Queries
If you parameterize the SQL string yourself, you can call the parameterizeQuery()
function
//connection poolvar redshiftClient = ; // options is an optional object with one property so far {raw: true} returns // just the data from redshift. {raw: false} returns the data with the pg objectredshiftClient;//you can also use promises to get the data
Template Literal Queries
If you use template literals to write your SQL, you can use a tagged template parser like https://github.com/felixfbecker/node-sql-template-strings to parameterize the template literal
//connection poolvar redshiftClient = ;var SQL = ; // options is an optional object with one property so far {raw: true} returns // just the data from redshift. {raw: false} returns the data with the pg objectlet value = 42; redshiftClient;//you can also use promises to get the data
rawQuery()
If you want to make a one time raw query, but you don't want to call connect & disconnect manually and you dont want to use conection pooling, you can use rawQuery()
//connection poolvar redshiftClient = ; // options is an optional object with one property so far {raw: true} returns // just the data from redshift. {raw: false} returns the data with the pg objectredshiftClient;//you can also use promises to get the data
Query Options
There's only a single query option so far. For the options object, the only valid option is {raw: true}, which returns just the data from redshift. {raw: false} or not specifying the value will return the data along with the entire pg object with data such as row count, table statistics etc.
CLI
There's a CLI with options for easy migration management. Creating a migration will create a redshift_migrations/
folder with a state file called .migrate
in it which contains the state of your completed migrations. The .migrate file keeps track of which migrations have been run, and when you run db:migrate, it computes the migrations that have not yet been run on your Redshift instance and runs them and saves the state of .migrate
WARNING!!! IF YOU HAVE SEPARATE DEV AND PROD REDSHIFT INSTANCES, DO NOT COMMIT THE .migrate
FILE TO YOUR VCS OR DEPLOY TO YOUR SERVERS. YOU'LL NEED A NEW VERSION OF THIS FILE FOR EVERY INSTANCE OF REDSHIFT.
Create a new migration file in redshift_migrations/ folder
node_modules/.bin/node-redshift migration:create <filename>
Run all remaining migrations on database
node_modules/.bin/node-redshift db:migrate <filename>
Undo last migration
node_modules/.bin/node-redshift db:migrate:undo <filename>
Creating a model using the command line
node_modules/.bin/node-redshift model:create <filename>
Models
A model will look like this
'use strict'; var person = 'tableName': 'people' 'tableProperties': 'id': 'type': 'key' 'name': 'type': 'string' 'required': true 'email': 'type': 'string' 'required': true ; moduleexports = person;
Importing and using model with ORM
There are two ways you could import and use redshift models. The first is using redshift.import in every file where you want to use the model ORM.
var redshift = ;var person = redshift; person;
The alternative(my preferred way) is to abstract the import calls and export all the models with the redshift object right after initialization
//redshift.js...redshift connection code... var person = redshift;redshiftmodels = {};redshiftmodelsperson = person; moduleexports = redshift; //usage in person.jsvar redshiftConnection = ;var person = redshiftmodelsperson; person;
ORM API
There are 3 functions supported by the ORM
/** * create a new instance of object * @param {Object or Array} data Object/Array with keys/values to create in database. keys are column names, values are data * @param * @return */Person; /** * update an existing item in redshift * @param * @param * @param * @return * */Person; /** * delete rows from redshift * @param * @param * @return */Person;
Upcoming features
- Ability to customize location of
.migrate
file or even from S3 - Model checking prior to queries to verify property name and type
- Add class & instance methods to model
License
MIT