MySQL: CHARSET from latin1 to utf8 > MySQL Tip

본문 바로가기
 

MySQL: CHARSET from latin1 to utf8

페이지 정보

작성자 no_profile 차동박 쪽지보내기 메일보내기 자기소개 아이디로 검색 전체게시물 댓글 0건 조회 21,009회 작성일 08-01-01 18:47

본문

Wednesday, March 29, 2006

MySQL: CHARSET from latin1 to utf8

A website im supporting needs to have multilingual characters. The default character set for MySQL is latin1. This unfortunately will not support Chinese nor other wierd multibyte characters.

It will quietly support them, but returns gibberish and will cause frustration all round.

After digging around, the best character set to use is UTF8.
To set the default charset for the server, the my.cfg/my.ini file has to be modified:

default-character-set=utf8


Unfortunately, once a database and their tables have been defined as latin1, they remain as latin1 unless you run this for each database:

alter database mydatabase charset=utf8;


and for each table:
alter table mytable charset=utf8;


and for each varchar/char type column:
alter table mytable alter column mycol charset=utf8;


and repeat ad infinitum....

This is rather tedious and boring, so there should be a better way. And that is to dump out the sql, change the charset and dump it back in. Here is the script.


===== latin1ToUTF8.sh

echo Script to convert MySQL latin1 charsets to utf8.
echo Usage: $0 dbname
echo 060329 yky Created.

echo Dumping out $1 database
mysqldump --add-drop-table $1 > db.sql

mydate=`date +%y%m%d`
echo Making a backup
mkdir bak &> /dev/null
cp db.sql bak/$1.$mydate.sql

echo String replacing latin1 with utf8
cat db.sql | replace CHARSET=latin1 CHARSET=utf8 > db2.sql

echo Pumping back $1 into database
mysql $1 < db2.sql

echo Changing db charset to utf8
mysql $1 -e "alter database $1 charset=utf8;"

echo $1 Done!


======

There must be a better way ?!

yk.

5 lewsers:

Simos said...
It's better to use "iconv" to convert the existing database to UTF-8. It's possible that the database already contains bad characters, so these have to be removed.
Once you have a sane UTF-8 database, you can import.

iconv -f utf-8 -t utf-8 < olddatabase.sql

To check if the database is sane.
Yoon Kit said...
Here are more resources:

http://blog.firebright.com/2006/09/12/mysql-charset-latin1-to-utf8-and-back-again-seeking-xen-with-character-encoding/

where he rates my post a mere 4 of 10.

And rightly so, because this one is the ultimate resource:

http://www.oreillynet.com/onlamp/blog/2006/01/turning_mysql_data_in_latin1_t.html


yk.
Alf said...
Here's a better way...it worked for me in MySQL 5, not sure about older versions:

alter table TABLE_NAME convert to character set utf8;
EoD said...
Try:
ALTER TABLE tbl_name CONVERT TO CHARACTER SET utf8;
(http://dev.mysql.com/doc/refman/5.0/en/alter-table.html)
eod said...
Oh, didn't mention alf.

Btw:
Converting from latin1 to utf8 caused some trouble with length of primary key (max 1000 byte) (since utf8 needs more space?)

댓글목록

등록된 댓글이 없습니다.

Total 243건 3 페이지
MySQL Tip 목록
번호 제목 글쓴이 조회 날짜
열람중 no_profile 차동박 쪽지보내기 메일보내기 자기소개 아이디로 검색 전체게시물 21010 01-01
206 no_profile 차동박 쪽지보내기 메일보내기 홈페이지 자기소개 아이디로 검색 전체게시물 20971 07-08
205 차동박 이름으로 검색 20960 08-07
204 차동박 이름으로 검색 20959 08-07
203 차동박 이름으로 검색 20833 08-07
202 차동박 이름으로 검색 20660 08-07
201 no_profile 차동박 쪽지보내기 메일보내기 홈페이지 자기소개 아이디로 검색 전체게시물 20374 05-06
200 no_profile 차동박 쪽지보내기 메일보내기 홈페이지 자기소개 아이디로 검색 전체게시물 20271 10-01
199 차동박 이름으로 검색 20253 12-27
198 no_profile 차동박 쪽지보내기 메일보내기 자기소개 아이디로 검색 전체게시물 20234 11-06
197 차동박 이름으로 검색 20220 07-20
196 차동박 이름으로 검색 20152 09-02
195 no_profile 차동박 쪽지보내기 메일보내기 홈페이지 자기소개 아이디로 검색 전체게시물 20115 01-12
194 차동박 이름으로 검색 20100 03-05
193 차동박 이름으로 검색 20080 08-07
192 차동박 이름으로 검색 19951 08-07
191 차동박 이름으로 검색 19788 08-07
190 차동박 이름으로 검색 19787 10-19
게시물 검색