|
laurent le cadet
|
Tuesday 23 August 2005 2:05:09 am
Hi, Here's one more message about db charset conversion but what I read previously poses me many questions ( http://ez.no/products/ez_publish_cms/documentation/configuration/configuration/language_and_charset/unicode_with_ez_publish ). I'm using eZp 3.5.2 revision 10972 with PHP 4.3.8 / MySQL 3.23.58, DB internal charset iso-8859-1. Actually there is a lot of content on the site (fre-FR + en-GB) and I want to add Chinese and Japanese. What I understand is to change the Charset to UTF-8. The previous threads about thid refered to mysql version problem or many different things to do... Actually, is there any step by step doc to perform this ? Regards. Laurent
|
|
Georg Franz
|
Tuesday 23 August 2005 6:27:51 am
Hi Laurent, 1st of all, you need a newer mysql, (the version should be greater or equal than 4.1.11). Then you need to change the charset of the ez tables. I've rewritten a small script for that purpous which I found at the forums of mysql:
<?php
// put in your username, password
$conn = mysql_connect("localhost", "root", "mypassword");
//change this to false to alter on the fly
$printonly=true;
$charset="utf8";
$collate="utf8_general_ci";
$altertablecharset=true;
$alterdatabasecharser=true;
// put here your databases ...
$currentDBArray = array();
$currentDBArray[] = "mydb";
function PMA_getDbCollation($db)
{
$sq='SHOW CREATE DATABASE `'.$db.'`;';
$res = mysql_query($sq);
if(!$res) echo "\n\n".$sq."\n".mysql_error()."\n\n"; else
if($row = mysql_fetch_assoc($res))
{
$tokenized = explode(' ', $row[1]);
unset($row, $res, $sql_query);
for ($i = 1; $i + 3 < count($tokenized); $i++)
{
if ($tokenized[$i] == 'DEFAULT' && $tokenized[$i + 1] == 'CHARACTER' && $tokenized[$i + 2] == 'SET')
{
if (isset($tokenized[$i + 5]) && $tokenized[$i + 4] == 'COLLATE')
{
return array($tokenized [$i + 3],$tokenized[$i + 5]); // We found the collation!
}
else
{
return array($tokenized [$i + 3]);
}
}
}
}
return '';
}
$rs2 = mysql_query("SHOW DATABASES");
if(!$rs2)
echo "\n\n".$sq."\n".mysql_error()."\n\n";
else
while ($data2 = mysql_fetch_row($rs2))
{
$db=$data2[0];
$db_cha=PMA_getDbCollation($db);
if ( in_array ( $db, $currentDBArray ) )
if ( substr($db_cha[0],0,4)!='utf8' ) // limit to charset
{
mysql_select_db($db);
$rs = mysql_query("SHOW TABLES");
if(!$rs)
echo "\n\n".$sq."\n".mysql_error()."\n\n";
else
while ($data = mysql_fetch_row($rs))
{
if ( substr ( $data[0], 0,2 ) == "ez" )
{
$rs1 = mysql_query("show FULL columns from $data[0]");
if(!$rs1)
echo "\n\n".$sq."\n".mysql_error()."\n\n";
else
while ($data1 = mysql_fetch_assoc($rs1))
{
if(in_array(array_shift(split("\\(",$data1['Type'],2)),array(
//'national char',
//'nchar',
//'national varchar',
//'nvarchar',
'char',
'varchar',
'tinytext',
'text',
'mediumtext',
'longtext',
'enum',
'set'
)))
{
if(substr($data1['Collation'],0,4)!='utf8') // limit to charset
{
$sq="ALTER TABLE `$data[0]` CHANGE `".$data1['Field'].'` `'.$data1['Field'].'` '.$data1['Type'].' CHARACTER SET binary '.($data1['Default']==''?'':($data1['Default']=='NULL'?' DEFAULT NULL':' DEFAULT \''.mysql_escape_string($data1['Default']).'\'')).($data1['Null']=='YES'?' NULL ':' NOT NULL').';';
if(!$printonly&&!mysql_query($sq)) echo "\n\n".$sq."\n".mysql_error()."\n\n";
else
{
echo ($sq."\n") ;
$sq="ALTER TABLE `$data[0]` CHANGE `".$data1['Field'].'` `'.$data1['Field'].'` '.$data1['Type']." CHARACTER SET $charset ".($collate==''?'':"COLLATE $collate").($data1['Default']==''?'':($data1['Default']=='NULL'?' DEFAULT NULL':' DEFAULT \''.mysql_escape_string($data1['Default']).'\'')).($data1['Null']=='YES'?' NULL ':' NOT NULL').($data1['Comment']==''?'':' COMMENT \''.mysql_escape_string($data1['Comment']).'\'').';';
if(!$printonly&&!mysql_query($sq)) echo "\n\n".$sq."\n".mysql_error()."\n\n";
else echo ($sq."\n") ;
}
}
}
}
if($altertablecharset)
{
/*
$sq='ALTER TABLE `'.$data[0]."` DEFAULT CHARACTER SET binary";
echo ($sq."\n") ;
if(!mysql_query($sq)) echo "\n\n".$sq."\n".mysql_error()."\n\n";
*/
$sq='ALTER TABLE `'.$data[0]."` DEFAULT CHARACTER SET $charset ".($collate==''?'':"COLLATE $collate");
echo ($sq."\n") ;
if(!$printonly)
if(!mysql_query($sq)) echo "\n\n".$sq."\n".mysql_error()."\n\n";
}
}
else
echo $data[0] . " nicht geändert.\n";
if( $alterdatabasecharser )
{
/*
$sq='ALTER DATABASE `'.$data2[0]."` DEFAULT CHARACTER SET binary";
echo ($sq."\n") ;
if(!mysql_query($sq)) echo "\n\n".$sq."\n".mysql_error()."\n\n";
*/
$sq='ALTER DATABASE `'.$data2[0]."` DEFAULT CHARACTER SET $charset ".($collate==''?'':"COLLATE $collate");
echo ($sq."\n") ;
if(!$printonly)
if(!mysql_query($sq)) echo "\n\n".$sq."\n".mysql_error()."\n\n";
}
}
}
}
?>
Then you need to change the ini-settings of ezpublish.
-> site.ini.append: charset at db entry
-> i18n.ini.append: charset-setting After that, don't forget to clear the ezpublish cache completly. HTH.
Best wishes,
Georg.
--
http://www.schicksal.com Horoskop website which uses eZ Publish since 2004
|
|
laurent le cadet
|
Tuesday 23 August 2005 8:42:30 am
Hi Georg,
Thanks for your repply.
First step : upgrade mysql...
After that, no risks for contents ?
Is the content of each table is re-encode ? (no need ?) About the script, I presume I just have to launch it one time from the root for the site (for example) ? Regards Laurent.
|
|
Georg Franz
|
Tuesday 23 August 2005 9:56:46 am
Hi Laurent, backup - backup - backup ... of course :-)) The script converts the tables first to a binary format and then to utf8, so no data should be lost. The script simply produces sql strings for the conversion. If you run it the first time and the var $printonly is set to true (at the begin of the script), only the sql strings are written to the screen, nothing else happen. If you really want to do the conversion, set $printonly to false.
Best wishes,
Georg.
--
http://www.schicksal.com Horoskop website which uses eZ Publish since 2004
|