id   numbers
1    2,34,647
2    78,102,342
3    34,78,103

Now our task would be to select IDs where numbers contain “2″ and “34″. One would say that we can do

SELECT * FROM mytable WHERE numbers LIKE "2" OR numbers LIKE "34"

But that won’t work correctly, because 2 will match 2, 102 and 342, which is something we don’t need. In this case MySQL word boundary regex comes handy:

SELECT * FROM mytable WHERE numbers REGEXP "[[:<:]]2[[:>:]]" OR numbers REGEXP "[[:<:]]34[[:>:]]"

Take a look at [[:<:]] and [[:>:]]. This is all we need here. The word boundary regex makes sure the match we are looking for is a standalone word, matching without commas and spaces, and making sure the “2″ word doesn’t have additional letters/numbers on left/right side. This is pretty awesome and saves our day!

This page can be found by searching for:

Коллекция стоковых фотографий

Which in my case was converted upon output to

Коллекция стоковых фотографий

This is ok if you need to just display it, but if you need to perform search or sort or any other mysql operations, you will FAIL. The fix for this is to convert the table column to binary, then back to UTF8. This way the data remains untouched but in proper encoding.

SET NAMES latin1; 
ALTER TABLE links MODIFY COLUMN link_title TEXT CHARACTER SET latin1; 
ALTER TABLE links MODIFY COLUMN link_title BLOB; 
ALTER TABLE links MODIFY COLUMN link_title TEXT CHARACTER SET utf8; 
SET NAMES utf8;

After that, don’t forget to always “SET NAMES utf8″ in your php application if you don’t have access to php.ini. It fixed my issues like magic and i continued working with same data without any loss.

This page can be found by searching for:

mysql_query("SET character_set_results = 'utf8', character_set_client = 'utf8', character_set_connection = 'utf8', character_set_database = 'utf8', character_set_server = 'utf8'");

Don’t forget to adjust the encoding names according to your configs (because not every website and database uses utf8).

Additionally, it may be possible to fix your encoding bugs with a simple single query (works in every 3rd situation):

mysql_set_charset('utf8');

This page can be found by searching for:

$sql = " 
CREATE TABLE IF NOT EXISTS `zcta` ( `zip` char(5) NOT NULL, `city` varchar(64) default NULL, PRIMARY KEY  (`zip`)) ENGINE=MyISAM DEFAULT CHARSET=latin1; 
INSERT INTO `zcta` (`zip`, `city`) VALUES
('00211', 'Portsmouth'),
('00212', 'Portsmouth');
";

Something like that, now we split the string where the ; is and make sure it isn’t the value inside the query (this weird regexp does it all)

$queries = preg_split("/;+(?=([^'|^\\\']*['|\\\'][^'|^\\\']*['|\\\'])*[^'|^\\\']*[^'|^\\\']$)/", $sql); 
foreach ($queries as $query){ 
   if (strlen(trim($query)) > 0) mysql_query($query); 
}

This page can be found by searching for:

To replace old url with new url in WordPress posts content:

UPDATE wp_posts SET post_content = REPLACE(post_content, "http://oldurl.com", "http://newurl.com")

To replace all <code> tags with <pre> run these 2 queries. Note that i put <code, that’s in case there are some attributes in the tag

UPDATE wp_posts SET post_content = REPLACE(post_content, "<code", "<pre")
UPDATE wp_posts SET post_content = REPLACE(post_content, "</code", "</pre")

This page can be found by searching for: