The first step is to fix the MySQL installation in order to store internationalized information., so locate your my.cnf configuration file on Linux, or the my.ini on Windows boxes.

Search in the configuration file the [mysqld] section when there is the configuration of the MySQL server.

Insert the following lines and eventually remove any existing configuration option with the same name.

[mysqld]
character-set-server=utf8
default-collation=utf8_unicode_ci

The option character-set-server=utf8 tells to the server that, if not otherwise specified, the character set of the created databases, tables, column will be utf8.

utf8 columns will be able to store cyrillic or simplified chinese character, just to give you two examples.

The collation defines how alphabetical ordering will happen, in few words which is the order of the letters that we expect on ORDER BY columnName clauses.

The suffix _ci means that ordering and comparison will be case insensitive and this is the common behavior used in databases.

Be very careful, because usually programming languages (i.e. Java) have case sensitive .equals(String string) method on String class, so it’s quite common to have some mistakes caused by this incongruency.

Then look for the [client] section of your configuration file, and write this line below it.

[client]
default-character-set=utf8

This is very important because it defines the character set used by the MySQL command-line client, and that’s what will be used to migrate the data from latin1 to utf8.

Now everything is setup, restart MySQL to make sure it’s using the updated configuration, and shut-down any application that is using the database that’s going to be migrated.

First, mysqldump will create a .sql file containing all the data:

mysqldump --skip-set-charset --no-create-db –no-create-info -h hostname --protocol=TCP -P 3306 -u username -p old_database > dump.sql

The option --skip-set-charset prevents that in the dump file will be any reference to the old (and wrong) character sets. The options --no-create-db and --no-create-info are used because the new database name will be defined later.

Now the new database is going to be created: mysql -u username -p and the following SQL should be executed in the terminal:

create schema new_database;
quit

Finally the last step is to populate the brand new database with the dumped data:

mysql -u username -p new_database < dump.sql

In this way all the previous data from old_database is now stored in utf8 format in new_database.

I hope this tutorial can be useful, please ask any question or give your feedback.
Thank You.