CakeDC Blog

TIPS, INSIGHTS AND THE LATEST FROM THE EXPERTS BEHIND CAKEPHP

Quick start with Migrations plugin

In a previous post I gave an overview of the CakePHP Migrations plugin, what it does and why you should use it in your applications. This article will explain how to use it in a practical way. We are going to bake a simple blog application recipe application and see how migrations are integrated in the development process.

Since we recently moved all our open source projects on http://cakedc.github.com/, this sample application source code is also available there: Sample Migrations Application - Github (it is a CakePHP 1.3 application). Ready?

Bake a new application and add the migrations plugin

First of all, we need to bake a new CakePHP application. Easy enough to do using cake bake, then configure your database (an empty database is sufficient for now) and check that the home page is all green! If you have not set up your environment to use the CakePHP command line yet, take some time to do so... it worth it!

Adding the migrations plugin might also be a straightforward task. You can either download the archive containing the plugin code and unzip it in the "/plugins/migrations" folder of your application, or  add it as a git submodule with the following command:

git submodule add git://github.com/CakeDC/Migrations.git plugins/migrations

Then check that it is correctly installed by executing the following command from your application root:

cake migration help

If you see a list of available commands you can move on next step.

Create initial tables and bake the MVC

We now need something to migrate! Let's create some tables in the database. The application will have Users who can publish Recipes, each one having several Ingredients (of course Ingredients can be used in many Recipes). Here is a SQL dump of this simple database schema:

CREATE TABLE `ingredients` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `name` varchar(100) NOT NULL,
  PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1;

CREATE TABLE `ingredients_recipes` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `ingredient_id` int(11) NOT NULL,
  `recipe_id` int(11) NOT NULL,
  PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1;

CREATE TABLE `recipes` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `user_id` int(11) NOT NULL,
  `name` varchar(100) NOT NULL,
  `content` text NOT NULL,
  `created` datetime NOT NULL,
  `modified` datetime NOT NULL,
  PRIMARY KEY (`id`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1;

CREATE TABLE `users` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `name` varchar(100) NOT NULL,
  `password` varchar(255) NOT NULL,
  `created` datetime NOT NULL,
  `modified` datetime NOT NULL,
  PRIMARY KEY (`id`)
) ENGINE=MyISAM  DEFAULT CHARSET=latin1;

As our goal here is not to focus on the application code itself, baked MVC from these tables might be sufficient... just run the command cake bake all for User, Recipe and Ingredient to bake'em all!

At this point we must have an application with an initial architecture ready to share. To start from here, one will just have to checkout the related commit... but don't you see a problem with this? How will he create the initial database? Maybe we could send him the SQL dump by email, or better commit it with the application! It is where the Migrations plugin comes in.

Generate the initial migration

"Be kind with your coworkers and include the database schema with your code... along with some sample data."

Let's use the migrations shell to generate an agnostic database schema containing our 4 tables, and an initial admin user account. To do so we just need to run the following command:

cake migration generate

After entering a name for the migration and selected the database dump option, we might have a new "/config/migrations" directory containing two files:

  • map.php representing the different migrations order,
  • name_of_the_migration.php a migration file containing all the necessary information to create your actual database. In the sample application it is named: "001_added_users_recipes_and_ingredients_tables.php". You might have noticed that we added a 001 prefix to the migration name to make it easier to see migrations order, it is a good practice.

We can now open the generated migration file (/config/migrations/001_added_users_recipes_and_ingredients_tables.php) and take a look at it. If you need more information and understand all available migration directives, you can read the plugin documentation.

For now we are just going to focus on the empty "after()" callback. This callback is triggered once the migration has been executed, and allow you to do whatever you want, given the direction of the migration: applied (up) or reverted (down). We are going to use this callback to create an initial admin User. Here is the code of the callback (as you are a CakePHP developer you might understand it quite easily):

function after($direction) {
	if ($direction === 'up') {
		if (!class_exists('Security')) {
			App::import('Core', 'Security');
		}

		$User = $this->generateModel('User');
		$user = array(
			'User' => array(
				'name' => 'admin',
				'password' => Security::hash('unsecurepassword', null, true)));
		$User->save($user);
	}
	return true;
}

Notice the use of the generateModel() method provided by the Migrations plugin. It is a shorthand allowing you to cleanly load a model in the callback to insert new data or update the existing. We could explain the reason of it more deeply but it is not the goal of this article, so just keep in mind that it is the best way to load a Model from callbacks!

Here we are! We can now share the application with anyone. After checked out the application, one will just have to run cake migration all to turn an empty database to a database containing all the needed tables, and an initial admin user to start using the application.

Categorize the recipes!

As the application evolves, we need to sort recipes by categories. This change involves two changes in the current database schema: a new categories table must be created, and a category_id field added to the recipes table.

Note: If you later want to use the migrations diff feature to generate a migration containing a diff between your previous database schema and the current one, you have to generate a Cake Schema of your database at this point. Simply run cake schema generate.

We can now update the recipes table and create a new categories table. Here is a simple SQL script:

CREATE TABLE `categories` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `name` varchar(100) NOT NULL,
  PRIMARY KEY (`id`)
) ENGINE=MyISAM  DEFAULT CHARSET=latin1;
ALTER TABLE `recipes` ADD `category_id` INT NOT NULL

Bake the MVC for categories and update recipes view pages to display the category so the application reflect these database changes. Before sharing these code changes, we need to generate a second migration describing the above SQL snippet in an agnostic way... and creating initial categories!

Nothing different than what we did previously: run cake migration generate, give a name to the migration, and choose between generating a diff from the schema.php file (if one was generated), generating a dump of the database (we will remove unnecessary instructions later) or generating an empty migration file. Once generated, it is always important to check the generated directives for the migration and fix them if needed. The migration must look like this:

var $migration = array(
	'up' => array(
		'create_table' => array(
			'categories' => array(
				'id' => array('type' => 'integer', 'null' => false, 'default' => NULL, 'key' => 'primary'),
				'name' => array('type' => 'string', 'null' => false, 'default' => NULL, 'length' => 100),
				'indexes' => array(
					'PRIMARY' => array('column' => 'id', 'unique' => 1),
				),
				'tableParameters' => array('charset' => 'latin1', 'collate' => 'latin1_swedish_ci', 'engine' => 'MyISAM'),
			),
		),
		'create_field' => array(
			'recipes' => array(
				'category_id' => array('type' => 'integer', 'null' => false, 'default' => NULL)
			),
		),
	),
	'down' => array(
		'drop_table' => array(
			'categories'
		),
		'drop_field' => array(
			'recipes' => array(
				'category_id'
			),
		),
	),
);

If you understood what we did in the first migration callback to add an initial user you might be able to implement this one. We would like to add initial categories: Starters, Main Dish and Desserts.

For lazy people, the code is here:

function after($direction) {
	if ($direction === 'up') {
		$Category = $this->generateModel('Category');
		$categories = array(
			array('name' => 'Starters'),
			array('name' => 'Main Dish'),
			array('name' => 'Desserts'));
		$Category->saveAll($categories);
	}
	return true;
}

Here we are again! The changes are ready to commit, and the commit will contains both code and database changes. One could update the database after checking out this commit by running: cake migration all.

The end

I hope this very simple use case and the code we built will help you to start using Migrations. As you could see it is very simple to use and will make your life much more easier: you would not have to worry anymore about the state of your database schema.

The source code of this tutorial is available on Github. If you found any bug or have any suggestion about the Migrations plugin, please create a ticket on Github. Comment this article if you have any question, and do not hesitate to share it if you found it useful!

Latest articles

Window functions

This article is part of the CakeDC Advent Calendar 2025 (December 15th 2025) Did you ever wanted to provide a partial result as part of an existing report? Window functions were added in CakePHP 4.1 and provide a way to pull a rolling result expressed naturally using the ORM. We'll use CakePHP 5 code in this article. Apart from the examples described in the book https://book.cakephp.org/5/en/orm/query-builder.html#window-functions One common scenario where window functions are very useful are rolling results. Imagine we have a transactions table, where account transactions are stored including a dollar amount of the transaction. The following migration would describe an example transactions table class CreateTransactions extends \Migrations\BaseMigration { public function change(): void { $table = $this->table('transactions'); $table ->addColumn('occurred_on', 'date', [ 'null' => false, ]) ->addColumn('debit_account', 'string', [ 'limit' => 255, 'null' => false, ]) ->addColumn('credit_account', 'string', [ 'limit' => 255, 'null' => false, ]) ->addColumn('amount_cents', 'biginteger', [ 'null' => false, 'signed' => false, ]) ->addColumn('currency', 'string', [ 'limit' => 3, 'null' => false, 'default' => 'USD', ]) ->addColumn('reference', 'string', [ 'limit' => 255, 'null' => true, ]) ->addColumn('description', 'string', [ 'limit' => 255, 'null' => true, ]) ->addTimestamps('created', 'modified') ->addIndex(['occurred_on'], ['name' => 'idx_transactions_occurred_on']) ->addIndex(['debit_account'], ['name' => 'idx_transactions_debit_account']) ->addIndex(['credit_account'], ['name' => 'idx_transactions_credit_account']) ->addIndex(['reference'], ['name' => 'idx_transactions_reference']) ->create(); } } Now, let's imagine we want to build a report to render the transaction amounts, but we also want a rolling total. Using a window function, we could define a custom finder like this one: public function findWindowReport( SelectQuery $query, ?string $account, ?Date $from, ?Date $to ): SelectQuery { $q = $query ->select([ 'id', 'occurred_on', 'debit_account', 'credit_account', 'amount_cents', 'currency', 'reference', 'description', ]); // Optional filters if ($account) { $q->where(['debit_account' => $account]); } if ($from) { $q->where(['occurred_on >=' => $from]); } if ($to) { $q->where(['occurred_on <=' => $to]); } $runningWin = (new WindowExpression()) ->partition('debit_account') ->orderBy([ 'occurred_on' => 'ASC', 'id' => 'ASC' ]); $q->window('running_win', $runningWin); $q->select([ 'running_total_cents' => $q ->func()->sum('amount_cents') ->over('running_win'), ]); return $q->orderBy([ 'debit_account' => 'ASC', 'occurred_on' => 'ASC', 'id' => 'ASC' ]); } Note the WindowExpression defined will sum the amount for each debit_account to produce the running_total_cents. The result of the report, after formatting will look like this Occurred On Debit Account Credit Account Amount (USD) Running Total (USD) 1/3/25 assets:bank:checking income:services $2,095.75 $2,095.75 1/3/25 assets:bank:checking income:sales $2,241.42 $4,337.17 1/7/25 assets:bank:checking income:services $467.53 $4,804.70 1/10/25 assets:bank:checking income:subscriptions $2,973.41 $7,778.11 1/12/25 assets:bank:checking income:sales $2,747.07 $10,525.18 1/17/25 assets:bank:checking income:subscriptions $2,790.36 $13,315.54 1/21/25 assets:bank:checking income:subscriptions $1,891.35 $15,206.89 1/28/25 assets:bank:checking equity:owner $353.00 $15,559.89 Other typical applications of window functions are leaderboards (building paginated rankins with scores, sales, activities), analytics for cumulative metrics (like inventory evolution) and comparison between rows (to compute deltas) and de-duplication (to pick the most recent record for example). This is a very useful tool to provide a solution for these cases, fully integrated into the CakePHP ORM. This article is part of the CakeDC Advent Calendar 2025 (December 15th 2025)

CounterCacheBehavior in CakePHP

This article is part of the CakeDC Advent Calendar 2025 (December 2th 2025)

CounterCacheBehavior in CakePHP: what it is, when to use it, and what’s new in CakePHP 5.2

As your application grows, a very common pattern appears: you need to display things like “number of comments”, “number of tasks”, or “number of orders”, and you need to do it fast. Calculating these values with COUNT() queries can work until performance starts to suffer (and complexity increases because of filters, states, or joins). This is exactly where CounterCacheBehavior* becomes useful.

What is CounterCacheBehavior?

CounterCacheBehavior is a CakePHP ORM behavior that keeps a counter field in a “parent” table synchronized based on the records in a related table. Typical example:
  • Articles hasMany Comments
  • You want to store the number of comments in articles.comment_count
The behavior automatically increments, decrements, or recalculates that value when related records are created, deleted, or modified.

When should you use it?

Common use cases include:
  • Listings with counters (e.g. “Posts (123 comments)”).
  • Sorting by counters (most commented, most active, etc.).
  • Filtering by counters (categories with more than X products).
  • Avoiding repeated and expensive COUNT( ) queries.
The idea is simple: accept a small cost on writes in exchange for much faster reads.

Basic configuration

CounterCache is configured in the child table (the one that belongs to the parent). If Comments belongsTo Articles, the behavior lives in CommentsTable. // src/Model/Table/CommentsTable.php namespace App\Model\Table; use Cake\ORM\Table; class CommentsTable extends Table { public function initialize(array $config): void { parent::initialize($config); $this->belongsTo('Articles'); $this->addBehavior('CounterCache', [ 'Articles' => ['comment_count'] ]); } } Doing this, CakePHP will automatically keep articles.comment_count up to date.

CounterCache with conditions (scoped counters)

Often you don’t want to count everything, but only a subset: published comments, active records, non-spam items, etc. $this->addBehavior('CounterCache', [ 'Articles' => [ 'published_comment_count' => [ 'conditions' => ['Comments.is_published' => true] ] ] ]); This pattern is very useful for dashboards such as:
  • open issues.
  • completed tasks.
  • approved records.

CounterCache with callbacks (custom calculations)

In some cases, conditions are not enough and you need more complex logic (joins, dynamic filters, or advanced queries). CounterCacheBehavior allows you to define a callable to calculate the counter value. Important: when using callbacks, bulk updates with updateCounterCache() will not update counters defined with closures. This is an important limitation to keep in mind.

What’s new in CakePHP 5.2: rebuild counters from the console

Before CakePHP 5.2, rebuilding counters often meant writing your own scripts or commands, especially after:
  • bulk imports done directly in the database.
  • manual data fixes.
  • adding a new counter cache in production.
  • data becoming out of sync.
New command: bashbin/cake counter_cache CakePHP 5.2 introduced an official command to rebuild counter caches: bin/cake counter_cache --assoc Comments Articles This command recalculates all counters related to Comments in the Articles table. Processing large tables in batches For large datasets, you can rebuild counters in chunks: bin/cake counter_cache --assoc Comments --limit 100 --page 2 Articles When using --limit and --page, records are processed ordered by the table’s primary key. This command is ideal for maintenance tasks and for safely backfilling new counter caches without custom tooling.

What’s new in CakePHP 5.2: bulk updates from the ORM

In addition to the console command, CakePHP 5.2 added a new ORM method: CounterCacheBehavior::updateCounterCache() This allows you to update counters programmatically, in batches: // Update all configured counter caches in batches $this->Comments->updateCounterCache(); // Update only a specific association, 200 records per batch $this->Comments->updateCounterCache('Articles', 200); // Update only the first page $this->Comments->updateCounterCache('Articles', page: 1); This is available since CakePHP 5.2.0.

Complete practical example: Articles and Comments

Assume the following database structure:
  • articles: id, title, comment_count (int, default 0), published_comment_count (int, default 0).
  • comments: id, article_id, body, is_published.

1) Behavior configuration in CommentsTable:

$this->addBehavior('CounterCache', [ 'Articles' => [ 'comment_count', 'published_comment_count' => [ 'conditions' => ['Comments.is_published' => true] ] ] ]);

2) Populate existing data (production)

After deploying, rebuild counters: bin/cake counter_cache --assoc Comments Articles From that point on, counters will stay synchronized automatically.

Best practices and Common Mistakes

Here you have some best practices and common mistakes:
  • Add indexes to foreign keys (comments.article_id) and fields used in conditions (comments.is_published) for large datasets.
  • If you perform direct database imports (bypassing the ORM), remember to rebuild counters using bin/cake counter_cache or updateCounterCache().
  • Counters defined using closures are not updated by updateCounterCache().
  • If a record changes its foreign key (e.g. moving a comment from one article to another), CounterCache handles the increments and decrements safely.
This article is part of the CakeDC Advent Calendar 2025 (December 2th 2025)

The Generational Perception of Work and Productivity in the Remote-Work Era

Generational Work Illustration

The Generational Perception of Work and Productivity in the Remote-Work Era

In the year 2020, everything changed when the world stopped completely during COVID-19. The perception of safety, health, mental health, work, and private life completely turned around and led to a different conception of the world we knew. As the global pandemic thrived, we saw how many jobs could be done from home, because people had to reinvent themselves as we were not able to go to our workplaces. And it settled a statement, changing the perception of work dramatically. Before it, and for older generations, work was associated with physical presence, rigid schedules, and productivity measured by visible hours. But after it, younger generations saw the potential of working from home or being a so-called digital nomad, giving more priority to flexibility, emotional well-being, and measuring efficiency through results. This change reflects a social evolution guided by new technologies, new expectations, and a more connected workforce. Remote work has been key in this transformation. For thousands of professionals, the ability to work from home meant reclaiming personal time, reducing stress, and achieving a healthier work--life balance (for example, by reducing commuting time most people get almost 2 extra hours of personal time). Productivity did not decrease --- in many cases, it actually improved --- because the focus shifted from "time spent" to "goals achieved." This model has also shown that trust and autonomy can lead to more engaged teams. However, despite all of the perks, many companies are apparently eager to return to traditional workplaces. Maybe it is the fear of losing control or a lack of understanding of the new work dynamics, but this tendency threatens to undo meaningful progress for generations that have already experienced the freedom and effectiveness of remote work. Going back to the old-fashioned way of work feels like a step backward. So now, the challenge is to find a middle ground that acknowledges the cultural and technological changes of our time, passing the torch to a new generation of workers. Because productivity is no longer measured by how many people are sitting in a chair, but by the value of the final results. And if we want organizations truly prepared for the future, we must listen to younger generations and build work models that prioritize both results and workers' well-being. In CakeDC we do believe in remote work! Proving through the years that work can be done remotely no matter the timezone or language.

We Bake with CakePHP