MySQL - 2

PHP 5.3 is released and after the release stress is over my mind is open for new ideas. While relaxing yesterday I thought about many things, among them was the Resultset iterator I recently discussed.

Now I wondered where to go next with this and had the idea that an individual Resultset is a child of the whole result and this might be wrapped in an Recursive Iterator. For doing so we don't implement the Iterator interface but RecursiveIterator. RecursiveIterator extends a typical Iterator with two methods: hasChildren() and getChildren(). But now we have a problem: The Iterator returned by getChildren() has to be a RecursiveIterator, too, which makes sense, in general. But I want to return a MySQLi Resultset which isn't recursive - so making this a RecursiveIterator is wrong. My solution now is to introduce yet another Iterator which goes by the name of MySQLi_PseudoRecursiveResultIterator and is implemented by extending IteratorIterator which will wrap the MySQLi_Result and implements RecursiveIterator telling the caller that there are no children.

As a sidenote: In our experimental tree Andrey made MySQLi_Result an iterator but that's not yet in php.net's CVS (might need some more testing, and probably we might change the design there...) so I'm emulating this with MySQLi_Result::fetch_all() combined with an ArrayIterator, using the experimental code the constructor can be dropped.

So let's finally look at the code of these two classes:

<?php
class MySQLi_ResultsetIterator implements RecursiveIterator
{
    private $mysqli;
    private $counter = 0;
    private $current = null;
    private $rewinded = false;

    public function __construct(mysqli $mysqli) {
        $this->mysqli = $mysqli;
    }
    private function freeCurrent() {
        if ($this->current) {
            $this->current->free();
            $this->current = null;
        }
    }
    public function rewind() {
        if ($this->rewinded) {
            throw new Exception("Already rewinded");
        }
        $this->freeCurrent();
        $this->counter = 0;
        $this->rewinded = true;
    }
    public function valid() {
        $this->current = $this->mysqli->store_result();
        return (bool)$this->current;
    }
    public function next() {
        $this->freeCurrent();
        $this->counter++;
        $this->mysqli->next_result();
    }
    public function key() {
        return $this->counter;
    }
    public function current() {
        if (!$this->current) {
            throw new Exception("valid() not called");
        }
        return $this->current;
    }

    public function hasChildren() {
        return true;
    }
    public function getChildren() {
        return new MySQLi_PseudoRecursiveResultIterator($this->current);
    }
}

class MySQLi_PseudoRecursiveResultIterator
     extends IteratorIterator
     implements RecursiveIterator
{
    public function __construct(MySQLi_Result $result) {
        // This ctor can be dropped with the experimental bzr sources
        // as IteratorIterator::__construct() directly works with
        // MySQLi_Result
        parent::__construct(new ArrayIterator($result->fetch_all()));
    }
    public function hasChildren() {
        return false;
    }
    public function getChildren() {
        throw new Exception("This should never be called");
    }    
}
?>

Now we can use this code. For properly using a RecursiveIterator one should use a RecursiveIteratorIterator (RII). To get some nice labels I'm extending the RII and then have a single foreach:

<?php
class MyRecursive_IteratorIterator
     extends RecursiveIteratorIterator
{
    public function __construct(MySQLi $mysqli, $flags = 0) {
        parent::__construct(
                new Mysqli_ResultSetIterator($mysqli),
                $flags | RecursiveIteratorIterator::LEAVES_ONLY);
    }

    public function beginChildren() {
        echo "Next ResultSet:\n";
    }
}

$mysqli = new MySQLi("localhost", "root", "", "test");

$query  = "SELECT 1,2 UNION SELECT 3, 4;".
          "SELECT 'hi world' UNION SELECT 'foobar'";
if ($mysqli->multi_query($query)) {
    foreach (new MyRecursive_IteratorIterator($mysqli) as $key => $row) {
        printf("    %s\n", $row[0]);
    }
}
?>

Now calling this code gives us a result similar to the following:

Next ResultSet:
    1
    3
Next ResultSet:
    hi world
    foobar

Isn't that nice? - I think that's a cool API! What do you think? Do you have use cases for such an API? Should we implement this in C and bundle it with PHP? Any feedback welcome!

Over at phpdeveloper.org I was pointed to a blog post talking about MySQLi and stored procedures. That reminded me about a small thing I recently did: When using MySQLi's multi_query to send queries which return multiple result sets you have to use a rather unintuitive API which can certainly be improved.

Recently I sat down and cooked up a small improvement for that, being an iterator fan I, of course, had to use an iterator for that and implemented the following class:

<?php
class MySQLi_ResultsetIterator implements Iterator {
    private $mysqli;
    private $counter = 0;
    private $current = null;
    private $rewinded = false;

    public function __construct(mysqli $mysqli) {
        $this->mysqli = $mysqli;
    }
    private function freeCurrent() {
        if ($this->current) {
            $this->current->free();
            $this->current = null;
        }
    }
    public function rewind() {
        if ($this->rewinded) {
            throw new Exception("Already rewinded, rewinding multiple times is not allowed!");
        }
        $this->freeCurrent();
        $this->counter = 0;
        $this->rewinded = true;
    }
    public function valid() {
        $this->current = $this->mysqli->store_result();
        return (bool)$this->current;
    }
    public function next() {
        $this->freeCurrent();
        $this->counter++;
        $this->mysqli->next_result();
    }
    public function key() {
        return $this->counter;
    }
    public function current() {
        if (!$this->current) {
            throw new Exception("valid() not called");
        }
?>

This iterator is wrapping all that's needed an then can be used like that:

<php
$mysqli = new MySQLi("localhost", "root", "", "test");

$query  = "SELECT 1,2 UNION SELECT 3, 4;".
          "SELECT 'hi world' UNION SELECT 'foobar'";
if ($mysqli->multi_query($query)) {
    foreach (new MySQLi_ResultsetIterator($mysqli) as $key => $result) {
        echo 'MySQL Resultset #'.(1+$key).":\n";
        while ($row = $result->fetch_row()) {
            printf("    %s\n", $row[0]);
        }
    }
}
?>

The output will be something like

MySQL Resultset #1:
    1
    3
MySQL Resultset #2:
    hi world
    foobar

And is, in my opinion, way nicer than the classical way, which you can see on the multi_query docs page.

That code is the first revision of that idea, I'll try to improve it and port it over to C so that some future version of PHP will include it. As a disclaimer: If you plan on using this class be aware that a future PHP might bundle a class having that exact name so use your own name Feedback welcome.

The time is just running and it's already time, again, for students to apply for this year's Google Summer of Code program. PHP and MySQL are among the list of 150 participating organizations. So if you're a student and are interested to learn how OpenSource works, do some networking with some famous people or just want a Google T-Shirt it's your time to take a look at the different idea pages (PHP, MySQL, others) or come up with an own idea and apply. Oh if accepted you even can earn some money as part of the program ...

Sometimes one has weird ideas, or am I the only one? - This specific one is at least a year old, now, during the Christmas days, waiting for New Year's Eve I had the time and mood to finally try it out: MySQL 5.1 has a plugin interface to easily add storage engines. PHP can easily embedded into other applications.  So why not combine these two things? - Writing a MySQL Storage Engine which reads data by calling a PHP script.

Let's start with a simple example first:

<?phpfunction create_table($table, $data) {
    return true;
}

function open_table($table) {
    return new ArrayIterator(array(
        array('id' => 1, 'dat' => 'foo'),
        array('id' => 2, 'a' => 'bar')
    ));
}
?>

This is the bare minimum storage engine my plugin supports. create_table() is called for creating the table, open_table() to access it, the later one then returns an iterator which is used for a full table scan. This example uses an ArrayIterator, which implements the SeekableIterator and the Countable interfaces, the first one provides a seek() method, which is called to read specific rows after sorting for instance, the later provides a method count() which gives the optimizer a hint.

Let's use this table:

mysql> CREATE TABLE php_test (id int, val CHAR(3)) ENGINE=PHP;
Query OK, 0 rows affected (0.04 sec)

mysql> SELECT * FROM php_test;
+------+------+
| id   | val  |
+------+------+
|    1 | foo  | 
|    2 | bar  | 
+------+------+
2 rows in set (0.00 sec)

Ok, of course that's nice and shiny but well, it's read only. To solve that you can implement a few interfaces provided by the plugin to handle writes:

<?php
class Test extends ArrayIterator 
implements MySQLStorage_Writable, MySQLStorage_Updatable, MySQLStorage_Deletable {
    public function write($data) {
        $this[] = $data;
    }

    public function update($data) {
        $this[$this->key()] = $data;
    }

    public function delete() {
        unset($this[$this->key()]);
    }
}

function create_table($table, $data) {
    return true;
}

function open_table($table) {
    return new Test(array(
        array('id' => 1, 'dat' => 'foo'),
        array('id' => 2, 'a' => 'bar')
    ));
}
?>

Again, we can test it:

mysql> UPDATE php_test SET val = 'baz' WHERE id = 1;
Query OK, 1 row affected (0.02 sec)
Rows matched: 1  Changed: 1  Warnings: 0

mysql> DELETE FROM php_test WHERE id = 2;
Query OK, 1 row affected (0.00 sec)

mysql> INSERT INTO php_test VALUES(3, 'bar');
Query OK, 1 row affected (0.00 sec)

mysql> SELECT * FROM php_write;
+------+------+
| id   | val  |
+------+------+
|    1 | baz  | 
|    3 | bar  | 
+------+------+
2 rows in set (0.00 sec)

As a reminder: This is calling PHP for all these operations.

So what might real life use cases be, once the major issues in the code are fixed? I have a few ideas like

• A live-logfile query tool, not sure that's really need, see the primitive Apache httpd access_log parser which is provided with the code as an example
• Combine it with the embedded MySQL server and use this storage engine for your unit tests, "mock tables" ...

Any other ideas? - Leave a comment

Oh, and like most MySQL stuff nowadays: There's a launchpad project for this plugin.

Ever wondered what your PHP application and MySQL actually do? An experimental mysqlnd branch will give you full access to the network communication stream. Using a custom PHP stream filter you can then intercept the communication ... but let's start at the beginning:

When talking about mysqlnd - the mysql native driver for PHP - we always mention the fact it's native in a way that we're, when possible, using PHP infrastructure. The most common example here is the memory management. By directly using PHP's memory we can avoid unnecessary copies of data from the MySQL Client Library's memory into PHP memory.

<?php
$mysqli = mysqli_connect("localhost", "root", "", "test");
$stream = mysqli_conn_to_stream($mysqli);
stream_filter_append($stream, "mysql.server", STREAM_FILTER_READ);
?>

But there's more what we're doing. We're also using PHP's stream abstraction layer. From a development perspective the benefit is that we're using a tested abstraction from different stream implementations by different operating systems instead of writing our own. But, again, there's more to it: We can export the communication stream to PHP userland. We hesitated about exporting it for some time as it can be quite dangerous and you might easily corrupt the client-server- communication.

As Ulf mentioned during his IPC talk I recently pushed a mysqlnd branch to launchpad which adds a userspace function to mysqli which returns a PHP stream for a connection. Using that stream you can now send your own requests to the server and wait for the response. That might be nice in a way, but I guess you most likely won't have use for that. PHP streams allow you to do more: PHP streams give you the possibility to add filters to a stream. These filters allow you to intercept packages which are sent or received , read them, change them or do whatever you like. A very simple filter can be found on the launchpad site, mentioned above. That filter simply prints the information after replacing unprintable (binary) characters by dots.

Once again: Just a small step, the next one is to decode the MySQL protocol. For that I've written a simple decoder for the MySQL protocol, not complete, but enough to give an idea. The script, including the decoder and some sample code using it, is, as a sample, part of the branch. When running you will get some output like

Query:
-> 0 59: QUERY: SELECT TABLE_SCHEMA FROM INFORMATION_SCHEMA.TABLES LIMIT 1
<- 1 1: DATA
<- 2 52: FIELD INFO
<- 3 5: EOF
<- 4 19: DATA
<- 5 5: EOF
Invalid Query:
-> 0 29: QUERY: ghdfjtgfdrs tztr ttgdszthtdr
<- 1 183: ERROR: 2000You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'ghdfjtgfdrs tztr ttgdszthtdr' at line 1
Prepare:
-> 0 61: PREPARE: SELECT TABLE_NAME FROM INFORMATION_SCHEMA.STATISTICS LIMIT 2
<- 1 12: OK
<- 2 52: FIELD INFO
<- 3 5: EOF
Ping:
-> 0 1: PING
<- 1 7: OK
Execute:
-> 0 11: EXECUTE
<- 1 1: DATA
<- 2 52: FIELD INFO
<- 3 5: EOF
<- 4 15: OK
<- 5 15: OK
<- 6 5: EOF
EOF
-> 0 5: CLOSE_STMT
-> 0 1: QUIT

As one can see: The protocol isn't fully decoded yet so this all might be extended but for me it served the purpose well enough. For making real use out of this we're thinking about exporting the protocol decoder which exists within mysqlnd to PHP userland.

What are your ideas for such a feature? - Sending different queries to different servers? Rewriting queries? Sharding? Replication? Easy scaling of your application while refactoring your application? Let us know!

Angelo recently showed an easy way to dump SQL queries using DTrace, while reading the articles I felt that some important information is missing: The name of the user executing the query and the selected database. So I sat down a few minutes and tried to collect that data.

For the database name I found a quite simple solution: It is passed as parameter to the check_user() function to MySQL so we can easily add a thread-local variable to keep that name. Simple script for that:

#!/usr/sbin/dtrace -s

#pragma D option quiet

pid$1::*check_user*:entry
{
    self->db = arg4 ? copyinstr(arg4) : "(no schema)";
}

pid$1::*dispatch_command*:entry
{
    printf("%s: %s\n", self->db, copyinstr(arg2));
}

Getting the username is a bit harder, for the PID-provider we need a function where it is passed as parameter, best would be if it was passed as char*, but there's no such function reliably called. (For instance there are functions for checking max connections per user where all we need is passed in a proper way, but it's only called when a limit was set) The only way would be to take the user_connect property of the THD object which is passed to dispatch_command and then access the username (and hostname). But getting that working from within DTrace is quite some work. I prepared some scripts doing this with simple C structures for the second part of my DTrace article series, which is ready in my head and is waiting to be typed, so in theory it should be possible, anybody wants to try?

The best solution, of course, would be to add proper probes into the MySQL server code which provide all that information.

In January Sun announced to buy MySQL, today the deal was closed and MySQL is now part of probably the worlds biggest Open Source company. For a full coverage visit mysql.com, Kaj's Blog or Sun's new product site. I'll continue with my tasks now.