Hibernate Localization Error

onur · **Joined:** Thu Jan 26, 2006 6:19 am **Posts:** 19

Guys I've got a problem with Hibernate.
For a table named Interest (POJO class name is also the same). All the sql statements created by hibernate is faulty.

The sql statement that should be created is something like this:
select id, .. .. from interests interest0 ...

Here "interest0" is the lowercase of Interest plus "0".

However, hibernate uses my regional settings during the lower case convertion. Unfortunately, in Turkish the lowercase of I is not i. its another character like "ı".

How can I make it work? How can I tell hibernate not to use my local setting and use English as default??

Thanks..

max · **Posted:** Mon Feb 13, 2006 5:31 am

implement you own NamingStrategy or override the table name specifically via the table attribute ?

onur · **Joined:** Thu Jan 26, 2006 6:19 am **Posts:** 19

max wrote:

implement you own NamingStrategy or override the table name specifically via the table attribute ?

How can I verride the table name specifically via the table attribute ?

baliukas · **Posted:** Mon Feb 13, 2006 5:34 am

http://javaalmanac.com/egs/java.util/SetDefLocale.html

max · **Posted:** Mon Feb 13, 2006 5:36 am

table="interest"

baliukas · **Posted:** Mon Feb 13, 2006 5:36 am

Code:

/** Useful constant for country.
     */
    static public final Locale UK = new Locale("en","GB","");

    /** Useful constant for country.
     */
    static public final Locale US = new Locale("en","US","");

    /** Useful constant for country.
     */
    static public final Locale CANADA = new Locale("en","CA","");

onur · **Joined:** Thu Jan 26, 2006 6:19 am **Posts:** 19

baliukas wrote:

http://javaalmanac.com/egs/java.util/SetDefLocale.html

if I change local to "ENUS", wont I lose the i18n of my web application?
then none of the resources will be read from my turkish resource.properties file, right?

baliukas · **Posted:** Mon Feb 13, 2006 5:41 am

Yes, you must not depend on default locale. Use way sugested by Max.

onur · **Joined:** Thu Jan 26, 2006 6:19 am **Posts:** 19

max wrote:

table="interest"

Code:

<?xml version="1.0"?>
<!DOCTYPE hibernate-mapping PUBLIC
"-//Hibernate/Hibernate Mapping DTD//EN"
"http://hibernate.sourceforge.net/hibernate-mapping-3.0.dtd" >
<hibernate-mapping package="com.salatta.dal">
    <class name="Invitation" table="invitations">
        <id name="id" type="integer" column="INV_ID">
            <generator class="org.hibernate.id.IncrementGenerator" />
        </id>
        <property name="name" column="NAME" type="string" not-null="false" length="25" />
        <property name="surname" column="SURNAME" type="string" not-null="false" length="25" />
        <property name="email" column="EMAIL" type="string" not-null="false" length="50" />
        <property name="invitationCode" column="INV_CODE" type="string" not-null="false" length="25" />
        <property name="invitationStatus" column="INV_STATUS" type="integer" not-null="false" length="11" />
        <many-to-one name="member" column="MEMBER_ID" class="Member" not-null="false"></many-to-one>
    </class>
</hibernate-mapping>

This is my hbm.xml file..
" <class name='Invitation' " part is used during the lowercase uperation
class name in written in lowercase.
When I change my class name to "Nvitation" it works.
But this is not a good solution.

If I overwrite NamingStrategy and tell hibernate to use invitation for the lowercase.. how can I inform hibernate that it should use the my CustomNamingStrategy class?[/code]

max · **Posted:** Mon Feb 13, 2006 6:28 am

hmm...this part is actually not controllable.

The "faulty" code is StringHelper.generateAliasRoot(String description) which just uses lowercase.

Report it as a bug in jira.

baliukas · **Posted:** Mon Feb 13, 2006 6:59 am

Try "<?xml version="1.0" encoding="utf-8"?>" "toLowercase" method transforms UNICODE string (it does not depend on locale settings), you need to read correct characters from XML first (it depends on locale settings).
Probably it is a bug in parser utf-8 is default.

baliukas · **Posted:** Mon Feb 13, 2006 7:48 am

Sorry, I was wrong. It is a flaw in UNICODE standard, it use same code for all 'I' http://www.unicode.net/reports/tr21/tr21-3.html and it depends on locale settings and SUN hacket it this way:
public String toLowerCase() {
return toLowerCase(Locale.getDefault());
}
So you will need custom strategy:
http://www.hibernate.org/hib_docs/v3/re ... ngstrategy.

max · **Posted:** Mon Feb 13, 2006 10:06 am

the namingstrategy does not influence the alias generation afaik. I thought it did until now ;)

baliukas · **Posted:** Mon Feb 13, 2006 10:40 am

Yes, it needs paramater. It looks like UNICODE decided to save space in mapping table :)