Core concepts and design principles.md

Core concepts and design principles

Database agnostic

maeva is database-agnostic. This means you should be able to run your maeva models against any database vendor - provided there is an existing maeva connector for that database vendor.

Being database-agnostic has inherent limitations. The most obvious ones are:

• Different id or key implementation
• Different type acceptance
• Different ways of reading/writing data

# Ids and keys

maeva ignores ids. Instead, it is the responsibility of the maeva connector to implement it. For example, mysql uses a numeric id it increments on each row insertion, when mongodb will use an object, ObjectID. Actually, even in mysql, you could decide not to use an id. A key based database for example will have no id - instead it will have unique keys paired to a value. maeva ignores ids. It is up to the connector of implementing it.

import * as data from 'maeva';
import mysql from 'maeva-msyql';
import mongodb from 'maeva-mongodb';

const mysqlConn = data.connect(mysql('mysql://localhost'));
const mongoConn = data.connect(mongodb('mongodb://localhost'));

const userModel = data.model(
  'users',
  {email: String}
);

const mysqlUser = await data.insertOne(
  'users',
  {email: '[email protected]'},
  {connection: mysqlConn}
);
// {id: 1, email: "joe@doe.com"}

const mongoUser = await data.insertOne(
  'users',
  {email: '[email protected]'},
  {connection: mongoConn}
);
// {_id: ObjectId("5a24c5f44082213b366f27f6"), email: "joe@doe.com"}

Yet, to make it easy on developers, we offer two methods to retrieve a document's id while staying database-agnostic:

data.getDocumentId(mysqlUser); // 1
data.getDocumentId(mongoUser); // ObjectId("5a24c5f44082213b366f27f6")

// You could also retrieve info about id from connector
const mysqlId = data.getId(mysqlConn); // {"name": "id", "type": Number}
const mongoId = data.getId(mongoConn); // {"name": "_id", "type": ObjectId}