Querying a Database from the Web
In any script used to access a database from the Web, you follow some basic steps:
- Check and filter data coming from the user.
- Set up a connection to the appropriate database.
- Query the database.
- Retrieve the results.
- Present the results back to the user.
These are the steps we followed in the script results.php, so now let’s go through each of them in turn.
Checking and Filtering Input Data
You begin the script by stripping any whitespace that the user might have inadvertently entered at the beginning or end of his search term. You do this by applying the function trim() to the value of $_POST['searchterm'] when giving it a shorter name:
$searchterm=trim($_POST['searchterm']);
The next step is to verify that the user has entered a search term and selected a search type. Note that you check whether the user entered a search term after trimming whitespace from the ends of $searchterm. If you arrange these lines in the opposite order, you could encounter situations in which a user’s search term is not empty and therefore does not create an error message; instead, it is all whitespace, so it is deleted by trim():
if (!$searchtype || !$searchterm) { echo "You have not entered search details. Please go back and try again."; exit; }
You check the $searchtype variable, even though in this case it’s coming from an HTML SELECT. You might ask why you should bother checking data that has to be filled in. It’s important to remember that there might be more than one interface to your database. For example, Amazon has many affiliates who use its search interface. Also, it’s sensible to screen data in case of any security problems that can arise because of users coming from different points of entry.
When you plan to use any data input by a user, you need to filter it appropriately for any control characters. As you might remember, in Chapter 4, “String Manipulation and Regular Expressions,” we described the functions addslashes(), stripslashes(), and get_magic_quotes_gpc(). You need to escape data when submitting any user input to a database such as MySQL.
In this case, you check the value of the get_magic_quotes_gpc() function. It tells you whether quoting is being done automatically. If it is not, you use addslashes() to escape the data:
if (!get_magic_quotes_gpc()) { $searchtype = addslashes($searchtype); $searchterm = addslashes($searchterm); }
You also use stripslashes() on the data coming back from the database. If the magic quotes feature is turned on, the data will have slashes in it when it comes back from the database, so you need to take them out.
Here you use the function htmlspecialchars() to encode characters that have special meanings in HTML. The current test data does not include any ampersands (&), less than (<), greater than (>), or double quotation mark (") symbols, but many fine book titles contain an ampersand. By using this function, you can eliminate future errors.
Setting Up a Connection
The PHP library for connecting to MySQL is called mysqli (the i stands for improved). When using the mysqli library in PHP, you can use either an object-oriented or procedural syntax.
You use the following line in the script to connect to the MySQL server:
@ $db = new mysqli('localhost', 'bookorama', 'bookorama123', 'books');
This line instantiates the mysqli class and creates a connection to host localhost with username bookorama, and password bookorama123. The connection is set up to use the database called books.
Using this object-oriented approach, you can now invoke methods on this object to access the database. If you prefer a procedural approach, mysqli allows for this, too. To connect in a procedural fashion, you use
@ $db = mysqli_connect('localhost', 'bookorama', 'bookorama123', 'books');
This function returns a resource rather than an object. This resource represents the connection to the database, and if you are using the procedural approach, you will need to pass this resource in to all the other mysqli functions. This is very similar to the way the file-handling functions, such as fopen(), work.
Most of the mysqli functions have an object-oriented interface and a procedural interface. Generally, the differences are that the procedural version function names start with mysqli_ and require you to pass in the resource handle you obtained from mysqli_connect(). Database connections are an exception to this rule because they can be made by the mysqli object’s constructor.
The result of your attempt at connection is worth checking because none of the rest of code will work without a valid database connection. You do this using the following code:
if (mysqli_connect_errno()) { echo 'Error: Could not connect to database. Please try again later.'; exit; }
(This code is the same for the object-oriented and procedural versions.) The mysqli_connect_errno() function returns an error number on error, or zero on success.
Note that when you connect to the database, you begin the line of code with the error suppression operator, @. This way, you can handle any errors gracefully. (This could also be done with exceptions, which we have not used in this simple example.)
Bear in mind that there is a limit to the number of MySQL connections that can exist at the same time. The MySQL parameter max_connections determines what this limit is. The purpose of this parameter and the related Apache parameter MaxClients is to tell the server to reject new connection requests instead of allowing machine resources to be completely used up at busy times or when software has crashed.
You can alter both of these parameters from their default values by editing the configuration files. To set MaxClients in Apache, edit the httpd.conf file on your system. To set max_connections for MySQL, edit the file my.conf.
Choosing a Database to Use
Remember that when you are using MySQL from a command-line interface, you need to tell it which database you plan to use with a command such as
use books;
You also need to do this when connecting from the Web. The database to use is specified as a parameter to the mysqli constructor or the mysqli_connect() function. If you want to change the default database, you can do so with the mysqli_select_db() function. It can be accessed as either
$db->select_db(dbname)
or as
mysqli_select_db(db_resource, db_name)
Here, you can see the similarity between the functions that we described before: The procedural version begins with mysqli_ and requires the extra database handle parameter.
Querying the Database
To actually perform the query, you can use the mysqli_query() function. Before doing this, however, it’s a good idea to set up the query you want to run:
$query = "select * from books where ".$searchtype." like '%".$searchterm."%'";
In this case, you search for the user-input value ($searchterm) in the field the user specified ($searchtype). Notice the use of like for matching rather than equal: it’s usually a good idea to be more tolerant in a database search.
You can now run the query:
$result = $db->query($query);
Or, if you want to use the procedural interface, you use
$result = mysqli_query($db, $query);
You pass in the query you want to run and, in the procedural interface, the database link (again, in this case $db).
The object-oriented version returns a result object; the procedural version returns a result resource. (This is similar to the way the connection functions work.) Either way, you store the result in a variable ($result) for later use. This function returns false on failure.
Retrieving the Query Results
A large variety of functions is available to break the results out of the result object or identifier in different ways. The result object or identifier is the key to accessing the rows returned by the query.
In this example, you counted the number of rows returned and also used the mysqli_fetch_assoc() function.
When you use the object-oriented approach, the number of rows returned is stored in the num_rows member of the result object, and you can access it as follows:
$num_results = $result->num_rows;
When you use a procedural approach, the function mysqli_num_rows() gives you the number of rows returned by the query. You should pass it the result identifier, like this:
$num_results = mysqli_num_rows($result);
It’s useful to know this if you plan to process or display the results, because you now know how many there are and can loop through them:
for ($i=0; $i <$num_results; $i++) { // process results }
In each iteration of this loop, you call $result->fetch_assoc() (or mysqli_fetch_assoc()). The loop does not execute if no rows are returned. This is a function that takes each row from the resultset and returns the row as an array, with each key an attribute name and each value the corresponding value in the array:
$row = $result->fetch_assoc();
Or you can use a procedural approach:
$row = mysqli_fetch_assoc($result);
Given the array $row, you can go through each field and display it appropriately, as shown in this example:
echo "<br />ISBN: "; echo stripslashes($row['isbn']);
As previously mentioned, you call stripslashes() to tidy up the value before displaying it.
Several variations can be used to get results from a result identifier. Instead of an array with named keys, you can retrieve the results in an enumerated array with mysqli_fetch_row(), as follows:
$row = $result->fetch_row($result);
or
$row = mysqli_fetch_row($result);
The attribute values are listed in each of the array values $row[0], $row[1], and so on. (The mysqli_fetch_array() function allows you to fetch a row as either or both kinds of array.)
You could also fetch a row into an object with the mysqli_fetch_object() function:
$row = $result->fetch_object();
or
$row = mysqli_fetch_object($result);
You can then access each of the attributes via $row->title, $row->author, and so on.
Disconnecting from the Database
You can free up your resultset by calling either
$result->free();
or
mysqli_free_result($result);
You can then use
$db->close();
or
mysqli_close($db);
to close a database connection. Using this command isn’t strictly necessary because the connection will be closed when a script finishes execution anyway.