1. Insert distinct column values into new secondary table.
2. Replace (update) existing table fields with proper foreign key for the related record in the secondary table.

This process can be time consuming if attempted manually. Luckily, MySQL 4.1+ allows for subqueries as well as the INSERT… SELECT… syntax.

The first step just requires running that insert select.

INSERT INTO contact_specialties (specialty)
SELECT DISTINCT specialty FROM `contacts` WHERE specialty != '' ORDER BY specialty;

Now that you have the distinct values in the second table, you can run an update query that uses a subquery to identify the proper foreign key for the old value.

UPDATE contacts SET specialty = (
SELECT id FROM contact_specialties WHERE contact_specialties.specialty=contacts.specialty);


This method assumes that you’re moving to a many-to-one relationship. But what if you need to add a third table to transform it further into a many-to-many relationship?

INSERT INTO contact_specialties_link (contact_id, specialty_id)
SELECT contacts.id, contact_specialties.id
FROM contacts
LEFT JOIN contact_specialties ON contact_specialties.specialty = contacts.specialty;

By default, both PHP and MySQL operate using a single timezone which is identified during installation and is quite often the same as the host server. Both tools allow you to specify a different timezone on the fly, and all following date operations work based off of that locale.

The primary goal is to preserve a base standard time so that we can easily convert between different timezones, without loosing the ability to convert again. Two choices emerged – storing a Unix Time stamp or a date string stored using GMT/UTC time.

The Unix Time stamp is a count of the number of seconds since January 1 1970 00:00:00 GMT.

The only difference between the two is that the time stamp is an integer so it makes no sense when read directly from the database. People don’t immediately understand that the date 1248988280 refers to 21:11:20 30-Jul-09 GMT. It makes more sense to me to store the dates in a format that’s easy to read, so we decided on storing GMT.

In the following demonstration, I show how to get the local time first and then convert that to GMT.

$base_time = time();

// print my local time
print date("H:i:s d-M-Y T", $base_time); // 14:49:49 30-Jul-2009 PDT

// Convert my local time to GMT
$gmt = gmdate("H:i:s d-M-Y T", $base_time);
print $gmt; // 21:49:49 30-Jul-2009 GMT

However, when developing the application it makes most sense to just generate the GMT date first. PHP has a great function for handling that called gmdate:

gmdate("H:i:s d-M-Y T");

At this point you can store that GMT format to your database. In MySQL, the supported date string format is the standard YYYY-MM-DD HH:MM:SS. The important thing is that you remember that the times are in GMT and will always need to be converted to local time for display.

Once you pull a record with the GMT date, you can then convert it to a specific timezone. The following example requires PHP 5.2 or above.

$cnv_date = new DateTime($gmt);
$cnv_date->setTimeZone(new DateTimeZone('America/New_York'));
print $cnv_date->format('H:i:s d-M-Y T'); // 17:49:49 30-Jul-2009 EDT

Now, we have a standard date string in GMT stored in the database, and we can easily convert it to another timezone without changing the entire server timezone.

I would recommend that you wrap this functionality into a basic object within your application. Although it’s not much work for each conversion, it would further reduce the amount of code and give you a central point of management.

However, when running a php script from the command line I was getting an error:

Warning: mysql_connect(): Can't connect to local MySQL server through socket '/var/mysql/mysql.sock' (2)

After spending some time thinking it was the code, I discovered that the mysql.sock file was missing.

Others on the net found the file elsewhere, like /private/tmp/mysql.sock or /tmp/mysql.sock but I had absolutely no .sock file anywhere.

Turns out, it did exist, but was named differently. According to the php.ini, the file was located at

/usr/local/mysql/run/mysql_socket

So, the solution was to create a symlink in /var/mysql to the proper file.

sudo ln -s /usr/local/mysql/run/mysql_socket /var/mysql/mysql.sock


$ vi /etc/my.cnf


Then add in:


ft_min_word_len=3


Or whatever minimum word length you’re looking for. Then, restart:


/etc/init.d/mysqld restart


Then rebuild your fulltext index.


ALTER TABLE `mytable` DROP INDEX `myfulltext_index`
ALTER TABLE `mytable` ADD FULLTEXT `fulltext` (
`myfieldname` ,
`anotherfieldname` ,
)


In order to get support for mysql as well as tools necessary for compiling, you need to install the apple developer tools. Check your original mac disc as you may already have a copy.

Anyway, on with our install:

Apache

First, download the latest version of the apache httpd web server.

./configure --prefix=/usr/local/apache --enable-module=so --enable-module=rewrite

Then, run the standard make, sudo make install.

MySQL

Then download the latest mysql source file. Make sure you download the source “Compressed GNU TAR archive (tar.gz)” as we’ll need to custom compile it.

However, PHP5 no longer includes mysql shared libraries needed during the configure process. To get these, download the source file for the mysql version you just installed. Once downloaded, decompress it and run the following:

./configure --prefix=/usr/local/mysql --with-unix-socket-path=/usr/local/mysql/run/mysql_socket --with-mysqld-user=mysql --with-comment --with-debug

make

sudo make install

sudo /usr/local/mysql/bin/mysql_install_db --force

sudo mkdir /usr/local/mysql/run
sudo mkdir /usr/local/mysql/data

sudo chgrp -R mysql /usr/local/mysql

sudo chown -R mysql /usr/local/mysql/run /usr/local/mysql/var /usr/local/mysql/data

Then start the server with:

sudo /usr/local/mysql/bin/mysqld_safe --user=mysql &

zlib

Then, install zlib if it’s not already installed on your system. Download zlib, and it should be a basic ./configure, make, sudo make install process.

Once zlib is installed, make sure you know the path as we’ll need it during the php configure process.

PHP

Next, download the latest php package. Decompress it, and now we’re ready to configure.

./configure --prefix=/usr/local/apache --enable-cli --with-mysql=/usr/local/mysql --with-zlib-dir=/usr/local --with-apxs2=/usr/local/apache/bin/apxs

Then, run make and sudo make install to complete the process.

Restart apache and test everything out. If php didn’t do it automatically, add in AddType application/x-httpd-php .php to your httpd.conf file – commonly located in /usr/local/apache/conf/httpd.conf.

Then, start your webserver.

sudo /usr/local/apache/bin/apachectl start

And finally, once everything is done you’ll want to setup your account to properly allow access to mysql and apache commands. Just run the following to add their paths to your bash profile:

echo 'export PATH=/usr/local/mysql/bin:$PATH' >> ~/.bash_profile
echo 'export PATH=/usr/local/apache/bin:$PATH' >> ~/.bash_profile

#!bin/sh

date=`date +%Y-%m-%d_%Hh%M`

cd /my/path/for/backups/

mysqldump -umy_user -p"mypass" my_database > $date.sql

tar -zcvf $date.tgz *.sql

# uncomment this line to import the sql into another database, for example as a mirror
#mysql -umy_user -p"mypass" mirror_database < $date.sql

rm *.sql

Then, just setup a cron job to run this at whichever interval you want this to run.

[mysql]
ALTER TABLE mytable DROP PRIMARY KEY;
[/mysql]

Then, I add a new one:

[mysql]
ALTER TABLE `mytable` ADD `id` INT NOT NULL AUTO_INCREMENT PRIMARY KEY FIRST ;
[/mysql]

[mysql]
ORDER BY RAND() * weight_fieldname
[/mysql]

Now, my possible weights from weight_fieldname are 5, 20, 100. 5 is the default weight for every record – 20 increases the chances of the record a bit and 100 pretty much sets the record as an exclusive result. Now, if you only have a single weight applied then setting it to 0 will reduce the weight of an item, however this is very difficult when you start having multiple weights.

One additional problem is that if you have a serious amount of records, using ORDER BY RAND() is going to be a performance hit.