A MySQL wrapping class is one of the most common applications that we use very often, and certainly there are lots of good places on the Web where we can look for database classes. One of my favorites is PEAR::DB, which provides a lot of room to work with different database systems, and offers a wealth of advanced features, powerful enough to keep you entertained for a long time. If you feel curious about it, please visit http://pear.php.net/package/DB and give it a try.

However, for the purposes of this article, we’ll work with our own MySQL abstraction class, so we can see how composition is properly carried out. Building such a class is a pretty straightforward process, which has been covered in some of my previous articles explaining aggregation in PHP. Thus, I guess the creation process shouldn’t pose any difficulties.

Now, armed with a decent background about PHP database classes, we might define our own version in the following way:

class MySQLConnector {

            var $conId; // connection identifier

            var $host; // MySQL host

            var $user; // MySQL username

            var $password; // MySQL password

            var $database; // MySQL database

            // constructor

            function MySQLConnector($options=array()){

                        // validate incoming parameters

                        if(count($options)>0){

                                   foreach($options as $parameter=>$value){

                                               (!empty($value))?$this->{$parameter}=$value:die('Invalid parameter '.$parameter);

                                   }

                                   // connect to MySQL

                                   $this->connectDB();

                        }

                        else {

                                   die('No connection parameters were provided');

                        }

            }

            // connect to MySQL server and select database

            function connectDB(){

                        if(!$this->conId=mysql_connect($this->host,$this->user,$this->password)){

                                    $this->isError('Error connecting to the server');

                        }

                        if(!mysql_select_db($this->database,$this->conId)){

                                    $this->isError('Error selecting database');

                        }

            }

            // perform query

            function performQuery($query){

                        if(!$this->result=mysql_query($query,$this->conId)){

                                   $this->isError('Error performing query '.$query);

                        }

                        // return new Result object

                        return new Result($this,$this->result);

            }

            // display errors

            function isError($errorMsg){

                        die($errorMsg.' '.mysql_error());

            }

}

As you can see, the “MySQLConnector” class is closely similar to that presented in previous articles. As usual, the constructor accepts the parameters required to connect to MySQL, that is, host, username, password and the selected database, in the form of an $options array. Then, each parameter is properly validated and finally the connection to the database server is performed by calling the private method “connectDB()”.

Besides the constructor, the class presents an “isError()” method to handle basically any error that occurred during the course of the MySQL operations, stopping the class with a die statement, and displaying a custom error message along with the error generated by the mysql_error() function.

However, there is a noticeable difference in the “performQuery()” method. Notice that it takes a SQL query, then performs it against the database table, but this time it returns a new instance of a “Result” object. Here’s the list of the relevant method:

function performQuery($query){

            if(!$this->result=mysql_query($query,$this->conId)){

                        $this->isError('Error performing query '.$query);

            }

            // return new Result object

            return new Result($this,$this->result);

}

Since this method is creating a new “Result” object, we have a clue about what’s happening here. We’ve established a relationship between objects, where the object “Result” composes the “MySQLConnector” object. Indeed, the process of composition becomes very understandable.

Also, maintaining our database wrapping class as a completely separate piece of code from the “Result” class, implies a immediate benefit, because we’re clearly delimiting the operations related to each class. It’s important to know that we’re not conceiving a class as a bunch of properties and methods that don’t keep a well-defined relationship to each other.

So far, so good. We have our “MySQLConnector” class, which nicely creates the “Result” object. But, what’s next? It’s time to look at the structure of the “Result” class, so we can see how it works. Just join me in the next explanation.