续集:在生产中更改模型架构

本文介绍了续集:在生产中更改模型架构的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！ 问题描述

我们正在使用orm sequelize.js ，并已定义了这样一个模型:

We're using the orm sequelize.js and have defined a model as such:

module.exports = function(sequelize, DataTypes) { var Source = sequelize.define('Source', { name: { type: DataTypes.STRING, allowNull: false, unique: true } }, { paranoid: true }); return Source; };

将其部署到生产环境并使用sequelize.sync同步到数据库.下一步，我们添加一个参数:

This is deployed to production and sync'd to the database using sequelize.sync. Next step, we add a parameter:

module.exports = function(sequelize, DataTypes) { var Source = sequelize.define('Source', { name: { type: DataTypes.STRING, allowNull: false, unique: true }, location: { type: DataTypes.STRING } }, { paranoid: true }); return Source; };

但是，当部署到生产环境sequelize.sync时，不会添加此新参数.这是因为sync执行以下操作:

However, when deploying to production sequelize.sync does not add this new parameter. This is because sync does a:

CREATE TABLE IF NOT EXISTS

如果表存在，则实际上不更新架构.这是在其文档中指出的.

And does not actually update the schema if the table exists. This is noted in their documentation.

唯一的选择似乎是{ force: true }，但是对于生产数据库来说这是不可行的.

The only option seems to be to { force: true }, however this is not okay for a production database.

有人知道在需要更改时如何正确更新架构吗?

Does anyone know how to properly update the schema when changes are necessary?

推荐答案

您要实施Sequelize迁移:

You want to implement Sequelize migrations:

docs.sequelizejs/manual/tutorial/migrations.html

这些功能使您可以在已知状态之间转换开发人员数据库，登台数据库和生产数据库.

These will enable you to transition developer, staging, and production databases between known states.