You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@openjpa.apache.org by "seba.wagner@gmail.com" <se...@gmail.com> on 2011/10/01 11:13:37 UTC
Problems with Oracle 11g and primary_key as rowtype IDENTITY
Hi,
when using Oracle 11g with openJPA I do get the following warning while the
tables are populated by openJPA:
"Oracle" does not support auto-assigne d column values. The column may not
behave as desired.
... and the effect is obviously that, as soon as you try to add a new column
with a null value for a primary key, Oracle will throw an Exception...
I can read from the Issue tracker:
https://issues.apache.org/jira/browse/OPENJPA-1090
But actually the JIRA Issue 1090 is about adding this warning, not about how
to solve it.
What is actually the solution to fix it?
Is it database related or do we need to change our code?
Thank you!
Sebastian
--
Sebastian Wagner
http://www.openmeetings.de
http://www.webbase-design.de
http://www.wagner-sebastian.com
seba.wagner@gmail.com
Re: Problems with Oracle 11g and primary_key as rowtype IDENTITY
Posted by Rick Curtis <cu...@gmail.com>.
Thanks for sharing the documentation, one comment for you though... Please
reconsider listing RuntimeUnenhancedClasses as a recommended configuration.
We don't recommend using that feature as it has numerous problems.
Thanks,
Rick
On Thu, Dec 1, 2011 at 3:19 AM, seba.wagner@gmail.com <seba.wagner@gmail.com
> wrote:
> We finally where able to use Oracle, we've put some docs online howto
> configure JPA in our Spring enabled configuration, maybe useful for
> others:
> http://code.google.com/p/openmeetings/wiki/Configuration_Oracle
>
> Sebastian
>
> 2011/11/2 seba.wagner@gmail.com <se...@gmail.com>:
> > Thanks Miłosz,
> >
> > I will try it out and come back with the results!
> >
> > Thanks
> > Sebastian
> >
> >
> > 2011/10/25 Miłosz Tylenda <mt...@o2.pl>
> >>
> >> Sebastian,
> >>
> >> For the sequence name too long problem, you could try setting the
> >> maxAutoAssignNameLength dictionary property to the maximum sequence name
> >> length in Oracle.
> >>
> >> You can also set the sequence name via the autoAssignSequenceName
> >> dictionary property but in this case you will rather need to create the
> >> sequence manually.
> >>
> >> If the above fails, there is an option to override
> >> DBDictionary.getGeneratedKeySequenceName or
> >> OracleDictionary.getCreateTableSQL methods.
> >>
> >> You might want to open a JIRA [1] issue for this problem.
> >>
> >> Let us know if this helps.
> >>
> >> Cheers,
> >> Milosz
> >>
> >> [1] https://issues.apache.org/jira/browse/OPENJPA
> >>
> >>
> >> > Or let me re-phrase my question:
> >> >
> >> > Is there a possibility to define an alternative name for the sequence
> >> > or a table when you are using the option:
> >> > <property name="openjpa.jdbc.DBDictionary"
> >> > value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
> >> >
> >> > Or is it really required at this point to rename the table's primary
> key
> >> > ??
> >> >
> >> > The error message Oracle throws is like that: "could not create
> >> > sequence because "appointmentcategory_seq" is too long name."
> >> >
> >> > I will also cross post this question to the Oracle support forums,
> >> > maybe they also have an idea how to workaround that.
> >> >
> >> > Btw: Using
> >> > <property name="openjpa.jdbc.SynchronizeMappings" value="buildSchema"
> />
> >> >
> >> > Also throws some error in our application using Oracle database...
> >> > So I always need to manually delete all tables and re-run the
> >> > application so that the tables are created (including sequences). Is
> >> > there a way of forcing openJPA to drop all tables? In hibernate you
> >> > would say "create" instead of "update". In openJPA I don't see any
> >> > alternative to "buildSchema". Is there a "createSchema" ?!
> >> > The documentation does not say so much about schema update /
> >> > forwarding.I don't see any other option that "buildSchema" mentioned
> >> > in the docs...
> >> >
> >> > Thanks
> >> > Sebastian
> >> >
> >> >
> >> >
> >> > 2011/10/22 seba.wagner@gmail.com <se...@gmail.com>:
> >> > > Hi,
> >> > >
> >> > > we are still trying to figure out what is going wrong with Oracle
> and
> >> > > openJPA.
> >> > >
> >> > > Our configuration now looks like that:
> >> > > <properties>
> >> > > <property
> >> > > name="openjpa.RuntimeUnenhancedClasses" value="unsupported"/>
> >> > > <property
> >> > > name="openjpa.ConnectionDriverName"
> >> > > value="oracle.jdbc.driver.OracleDriver"/>
> >> > > <property name="openjpa.ConnectionProperties"
> >> > > value="username=SYSTEM, password=admin"/>
> >> > > <property name="openjpa.ConnectionURL"
> >> > > value="jdbc:oracle:thin:@localhost:1521:openmeetings"/>
> >> > > <property
> >> > > name="openjpa.jdbc.SynchronizeMappings" value="buildSchema"/>
> >> > > <property name="openjpa.Log"
> >> > > value="DefaultLevel=WARN, Tool=INFO" />
> >> > > <property name="openjpa.DataCache"
> >> > > value="true"/>
> >> > > <property name="openjpa.QueryCache"
> >> > > value="true"/>
> >> > > <property
> >> > > name="openjpa.jdbc.DBDictionary"
> >> > > value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
> >> > > <property
> >> > > name="openjpa.jdbc.QuerySQLCache" value="false"/>
> >> > > </properties>
> >> > >
> >> > > UseTriggersForAutoAssign=true is the magic key to use sequences as
> it
> >> > > seems ...
> >> > > however ...
> >> > >
> >> > > => The problem now is that Oracle complains that some sequences
> have a
> >> > > too long column name. How can we bypass that?
> >> > >
> >> > >
> >> > > Thanks
> >> > > Sebastian
> >> > >
> >> > >
> >> > >
> >> > > 2011/10/2 seba.wagner@gmail.com <se...@gmail.com>:
> >> > >> The guys from Oracle do (a bit offensive) argue that for such
> reasons
> >> > >> sequences do exist.
> >> > >>
> >> > >>
> >> > >>
> https://forums.oracle.com/forums/thread.jspa?messageID=9909360#9909360
> >> > >>
> >> > >> Quote:
> >> > >> "Oracle has sequences for this purpose.
> >> > >> They are either directly used in an INSERT statement, or assigned
> to
> >> > >> a
> >> > >> column by means of a trigger.."
> >> > >>
> >> > >> So my question might be: Why does openJPA not create sequences? Am
> I
> >> > >> missing
> >> > >> some param in my config?
> >> > >> I have seen for MySQL there is an additional "sequence-table"
> created
> >> > >> with
> >> > >> all index. In Oracle there seems to be no such table ... while
> except
> >> > >> the
> >> > >> persistance.xml the code is 1:1 the same.
> >> > >>
> >> > >> Thanks,
> >> > >> Sebastian
> >> > >>
> >> > >> 2011/10/1 seba.wagner@gmail.com <se...@gmail.com>
> >> > >>>
> >> > >>> Hi,
> >> > >>>
> >> > >>> when using Oracle 11g with openJPA I do get the following warning
> >> > >>> while
> >> > >>> the tables are populated by openJPA:
> >> > >>>
> >> > >>> "Oracle" does not support auto-assigne d column values. The
> column
> >> > >>> may
> >> > >>> not behave as desired.
> >> > >>>
> >> > >>> ... and the effect is obviously that, as soon as you try to add a
> >> > >>> new
> >> > >>> column with a null value for a primary key, Oracle will throw an
> >> > >>> Exception...
> >> > >>>
> >> > >>> I can read from the Issue tracker:
> >> > >>> https://issues.apache.org/jira/browse/OPENJPA-1090
> >> > >>>
> >> > >>> But actually the JIRA Issue 1090 is about adding this warning, not
> >> > >>> about
> >> > >>> how to solve it.
> >> > >>>
> >> > >>> What is actually the solution to fix it?
> >> > >>> Is it database related or do we need to change our code?
> >> > >>>
> >> > >>> Thank you!
> >> > >>> Sebastian
> >> > >>>
> >> > >>> --
> >> > >>> Sebastian Wagner
> >> > >>> http://www.openmeetings.de
> >> > >>> http://www.webbase-design.de
> >> > >>> http://www.wagner-sebastian.com
> >> > >>> seba.wagner@gmail.com
> >> > >>
> >> > >>
> >> > >>
> >> > >> --
> >> > >> Sebastian Wagner
> >> > >> http://www.openmeetings.de
> >> > >> http://www.webbase-design.de
> >> > >> http://www.wagner-sebastian.com
> >> > >> seba.wagner@gmail.com
> >> > >>
> >> > >
> >> > >
> >> > >
> >> > > --
> >> > > Sebastian Wagner
> >> > > http://www.openmeetings.de
> >> > > http://www.webbase-design.de
> >> > > http://www.wagner-sebastian.com
> >> > > seba.wagner@gmail.com
> >> > >
> >> >
> >> >
> >> >
> >> >
> >
> >
> >
> >
> > --
> > Sebastian Wagner
> > http://www.openmeetings.de
> > http://www.webbase-design.de
> > http://www.wagner-sebastian.com
> > seba.wagner@gmail.com
>
>
>
> --
> Sebastian Wagner
> http://www.openmeetings.de
> http://www.webbase-design.de
> http://www.wagner-sebastian.com
> seba.wagner@gmail.com
>
--
*Rick Curtis*
Re: Problems with Oracle 11g and primary_key as rowtype IDENTITY
Posted by "seba.wagner@gmail.com" <se...@gmail.com>.
We finally where able to use Oracle, we've put some docs online howto
configure JPA in our Spring enabled configuration, maybe useful for
others:
http://code.google.com/p/openmeetings/wiki/Configuration_Oracle
Sebastian
2011/11/2 seba.wagner@gmail.com <se...@gmail.com>:
> Thanks Miłosz,
>
> I will try it out and come back with the results!
>
> Thanks
> Sebastian
>
>
> 2011/10/25 Miłosz Tylenda <mt...@o2.pl>
>>
>> Sebastian,
>>
>> For the sequence name too long problem, you could try setting the
>> maxAutoAssignNameLength dictionary property to the maximum sequence name
>> length in Oracle.
>>
>> You can also set the sequence name via the autoAssignSequenceName
>> dictionary property but in this case you will rather need to create the
>> sequence manually.
>>
>> If the above fails, there is an option to override
>> DBDictionary.getGeneratedKeySequenceName or
>> OracleDictionary.getCreateTableSQL methods.
>>
>> You might want to open a JIRA [1] issue for this problem.
>>
>> Let us know if this helps.
>>
>> Cheers,
>> Milosz
>>
>> [1] https://issues.apache.org/jira/browse/OPENJPA
>>
>>
>> > Or let me re-phrase my question:
>> >
>> > Is there a possibility to define an alternative name for the sequence
>> > or a table when you are using the option:
>> > <property name="openjpa.jdbc.DBDictionary"
>> > value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
>> >
>> > Or is it really required at this point to rename the table's primary key
>> > ??
>> >
>> > The error message Oracle throws is like that: "could not create
>> > sequence because "appointmentcategory_seq" is too long name."
>> >
>> > I will also cross post this question to the Oracle support forums,
>> > maybe they also have an idea how to workaround that.
>> >
>> > Btw: Using
>> > <property name="openjpa.jdbc.SynchronizeMappings" value="buildSchema" />
>> >
>> > Also throws some error in our application using Oracle database...
>> > So I always need to manually delete all tables and re-run the
>> > application so that the tables are created (including sequences). Is
>> > there a way of forcing openJPA to drop all tables? In hibernate you
>> > would say "create" instead of "update". In openJPA I don't see any
>> > alternative to "buildSchema". Is there a "createSchema" ?!
>> > The documentation does not say so much about schema update /
>> > forwarding.I don't see any other option that "buildSchema" mentioned
>> > in the docs...
>> >
>> > Thanks
>> > Sebastian
>> >
>> >
>> >
>> > 2011/10/22 seba.wagner@gmail.com <se...@gmail.com>:
>> > > Hi,
>> > >
>> > > we are still trying to figure out what is going wrong with Oracle and
>> > > openJPA.
>> > >
>> > > Our configuration now looks like that:
>> > > <properties>
>> > > <property
>> > > name="openjpa.RuntimeUnenhancedClasses" value="unsupported"/>
>> > > <property
>> > > name="openjpa.ConnectionDriverName"
>> > > value="oracle.jdbc.driver.OracleDriver"/>
>> > > <property name="openjpa.ConnectionProperties"
>> > > value="username=SYSTEM, password=admin"/>
>> > > <property name="openjpa.ConnectionURL"
>> > > value="jdbc:oracle:thin:@localhost:1521:openmeetings"/>
>> > > <property
>> > > name="openjpa.jdbc.SynchronizeMappings" value="buildSchema"/>
>> > > <property name="openjpa.Log"
>> > > value="DefaultLevel=WARN, Tool=INFO" />
>> > > <property name="openjpa.DataCache"
>> > > value="true"/>
>> > > <property name="openjpa.QueryCache"
>> > > value="true"/>
>> > > <property
>> > > name="openjpa.jdbc.DBDictionary"
>> > > value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
>> > > <property
>> > > name="openjpa.jdbc.QuerySQLCache" value="false"/>
>> > > </properties>
>> > >
>> > > UseTriggersForAutoAssign=true is the magic key to use sequences as it
>> > > seems ...
>> > > however ...
>> > >
>> > > => The problem now is that Oracle complains that some sequences have a
>> > > too long column name. How can we bypass that?
>> > >
>> > >
>> > > Thanks
>> > > Sebastian
>> > >
>> > >
>> > >
>> > > 2011/10/2 seba.wagner@gmail.com <se...@gmail.com>:
>> > >> The guys from Oracle do (a bit offensive) argue that for such reasons
>> > >> sequences do exist.
>> > >>
>> > >>
>> > >> https://forums.oracle.com/forums/thread.jspa?messageID=9909360#9909360
>> > >>
>> > >> Quote:
>> > >> "Oracle has sequences for this purpose.
>> > >> They are either directly used in an INSERT statement, or assigned to
>> > >> a
>> > >> column by means of a trigger.."
>> > >>
>> > >> So my question might be: Why does openJPA not create sequences? Am I
>> > >> missing
>> > >> some param in my config?
>> > >> I have seen for MySQL there is an additional "sequence-table" created
>> > >> with
>> > >> all index. In Oracle there seems to be no such table ... while except
>> > >> the
>> > >> persistance.xml the code is 1:1 the same.
>> > >>
>> > >> Thanks,
>> > >> Sebastian
>> > >>
>> > >> 2011/10/1 seba.wagner@gmail.com <se...@gmail.com>
>> > >>>
>> > >>> Hi,
>> > >>>
>> > >>> when using Oracle 11g with openJPA I do get the following warning
>> > >>> while
>> > >>> the tables are populated by openJPA:
>> > >>>
>> > >>> "Oracle" does not support auto-assigne d column values. The column
>> > >>> may
>> > >>> not behave as desired.
>> > >>>
>> > >>> ... and the effect is obviously that, as soon as you try to add a
>> > >>> new
>> > >>> column with a null value for a primary key, Oracle will throw an
>> > >>> Exception...
>> > >>>
>> > >>> I can read from the Issue tracker:
>> > >>> https://issues.apache.org/jira/browse/OPENJPA-1090
>> > >>>
>> > >>> But actually the JIRA Issue 1090 is about adding this warning, not
>> > >>> about
>> > >>> how to solve it.
>> > >>>
>> > >>> What is actually the solution to fix it?
>> > >>> Is it database related or do we need to change our code?
>> > >>>
>> > >>> Thank you!
>> > >>> Sebastian
>> > >>>
>> > >>> --
>> > >>> Sebastian Wagner
>> > >>> http://www.openmeetings.de
>> > >>> http://www.webbase-design.de
>> > >>> http://www.wagner-sebastian.com
>> > >>> seba.wagner@gmail.com
>> > >>
>> > >>
>> > >>
>> > >> --
>> > >> Sebastian Wagner
>> > >> http://www.openmeetings.de
>> > >> http://www.webbase-design.de
>> > >> http://www.wagner-sebastian.com
>> > >> seba.wagner@gmail.com
>> > >>
>> > >
>> > >
>> > >
>> > > --
>> > > Sebastian Wagner
>> > > http://www.openmeetings.de
>> > > http://www.webbase-design.de
>> > > http://www.wagner-sebastian.com
>> > > seba.wagner@gmail.com
>> > >
>> >
>> >
>> >
>> >
>
>
>
>
> --
> Sebastian Wagner
> http://www.openmeetings.de
> http://www.webbase-design.de
> http://www.wagner-sebastian.com
> seba.wagner@gmail.com
--
Sebastian Wagner
http://www.openmeetings.de
http://www.webbase-design.de
http://www.wagner-sebastian.com
seba.wagner@gmail.com
Re: Problems with Oracle 11g and primary_key as rowtype IDENTITY
Posted by "seba.wagner@gmail.com" <se...@gmail.com>.
Thanks Miłosz,
I will try it out and come back with the results!
Thanks
Sebastian
2011/10/25 Miłosz Tylenda <mt...@o2.pl>
> Sebastian,
>
> For the sequence name too long problem, you could try setting the
> maxAutoAssignNameLength dictionary property to the maximum sequence name
> length in Oracle.
>
> You can also set the sequence name via the autoAssignSequenceName
> dictionary property but in this case you will rather need to create the
> sequence manually.
>
> If the above fails, there is an option to override
> DBDictionary.getGeneratedKeySequenceName or
> OracleDictionary.getCreateTableSQL methods.
>
> You might want to open a JIRA [1] issue for this problem.
>
> Let us know if this helps.
>
> Cheers,
> Milosz
>
> [1] https://issues.apache.org/jira/browse/OPENJPA
>
>
> > Or let me re-phrase my question:
> >
> > Is there a possibility to define an alternative name for the sequence
> > or a table when you are using the option:
> > <property name="openjpa.jdbc.DBDictionary"
> > value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
> >
> > Or is it really required at this point to rename the table's primary key
> ??
> >
> > The error message Oracle throws is like that: "could not create
> > sequence because "appointmentcategory_seq" is too long name."
> >
> > I will also cross post this question to the Oracle support forums,
> > maybe they also have an idea how to workaround that.
> >
> > Btw: Using
> > <property name="openjpa.jdbc.SynchronizeMappings" value="buildSchema" />
> >
> > Also throws some error in our application using Oracle database...
> > So I always need to manually delete all tables and re-run the
> > application so that the tables are created (including sequences). Is
> > there a way of forcing openJPA to drop all tables? In hibernate you
> > would say "create" instead of "update". In openJPA I don't see any
> > alternative to "buildSchema". Is there a "createSchema" ?!
> > The documentation does not say so much about schema update /
> > forwarding.I don't see any other option that "buildSchema" mentioned
> > in the docs...
> >
> > Thanks
> > Sebastian
> >
> >
> >
> > 2011/10/22 seba.wagner@gmail.com <se...@gmail.com>:
> > > Hi,
> > >
> > > we are still trying to figure out what is going wrong with Oracle and
> openJPA.
> > >
> > > Our configuration now looks like that:
> > > <properties>
> > > <property
> name="openjpa.RuntimeUnenhancedClasses" value="unsupported"/>
> > > <property
> name="openjpa.ConnectionDriverName"
> > > value="oracle.jdbc.driver.OracleDriver"/>
> > > <property name="openjpa.ConnectionProperties"
> > > value="username=SYSTEM, password=admin"/>
> > > <property name="openjpa.ConnectionURL"
> > > value="jdbc:oracle:thin:@localhost:1521:openmeetings"/>
> > > <property
> name="openjpa.jdbc.SynchronizeMappings" value="buildSchema"/>
> > > <property name="openjpa.Log"
> value="DefaultLevel=WARN, Tool=INFO" />
> > > <property name="openjpa.DataCache"
> value="true"/>
> > > <property name="openjpa.QueryCache"
> value="true"/>
> > > <property
> name="openjpa.jdbc.DBDictionary"
> > > value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
> > > <property
> name="openjpa.jdbc.QuerySQLCache" value="false"/>
> > > </properties>
> > >
> > > UseTriggersForAutoAssign=true is the magic key to use sequences as it
> seems ...
> > > however ...
> > >
> > > => The problem now is that Oracle complains that some sequences have a
> > > too long column name. How can we bypass that?
> > >
> > >
> > > Thanks
> > > Sebastian
> > >
> > >
> > >
> > > 2011/10/2 seba.wagner@gmail.com <se...@gmail.com>:
> > >> The guys from Oracle do (a bit offensive) argue that for such reasons
> > >> sequences do exist.
> > >>
> > >>
> https://forums.oracle.com/forums/thread.jspa?messageID=9909360#9909360
> > >>
> > >> Quote:
> > >> "Oracle has sequences for this purpose.
> > >> They are either directly used in an INSERT statement, or assigned to a
> > >> column by means of a trigger.."
> > >>
> > >> So my question might be: Why does openJPA not create sequences? Am I
> missing
> > >> some param in my config?
> > >> I have seen for MySQL there is an additional "sequence-table" created
> with
> > >> all index. In Oracle there seems to be no such table ... while except
> the
> > >> persistance.xml the code is 1:1 the same.
> > >>
> > >> Thanks,
> > >> Sebastian
> > >>
> > >> 2011/10/1 seba.wagner@gmail.com <se...@gmail.com>
> > >>>
> > >>> Hi,
> > >>>
> > >>> when using Oracle 11g with openJPA I do get the following warning
> while
> > >>> the tables are populated by openJPA:
> > >>>
> > >>> "Oracle" does not support auto-assigne d column values. The column
> may
> > >>> not behave as desired.
> > >>>
> > >>> ... and the effect is obviously that, as soon as you try to add a new
> > >>> column with a null value for a primary key, Oracle will throw an
> > >>> Exception...
> > >>>
> > >>> I can read from the Issue tracker:
> > >>> https://issues.apache.org/jira/browse/OPENJPA-1090
> > >>>
> > >>> But actually the JIRA Issue 1090 is about adding this warning, not
> about
> > >>> how to solve it.
> > >>>
> > >>> What is actually the solution to fix it?
> > >>> Is it database related or do we need to change our code?
> > >>>
> > >>> Thank you!
> > >>> Sebastian
> > >>>
> > >>> --
> > >>> Sebastian Wagner
> > >>> http://www.openmeetings.de
> > >>> http://www.webbase-design.de
> > >>> http://www.wagner-sebastian.com
> > >>> seba.wagner@gmail.com
> > >>
> > >>
> > >>
> > >> --
> > >> Sebastian Wagner
> > >> http://www.openmeetings.de
> > >> http://www.webbase-design.de
> > >> http://www.wagner-sebastian.com
> > >> seba.wagner@gmail.com
> > >>
> > >
> > >
> > >
> > > --
> > > Sebastian Wagner
> > > http://www.openmeetings.de
> > > http://www.webbase-design.de
> > > http://www.wagner-sebastian.com
> > > seba.wagner@gmail.com
> > >
> >
> >
> >
> >
>
--
Sebastian Wagner
http://www.openmeetings.de
http://www.webbase-design.de
http://www.wagner-sebastian.com
seba.wagner@gmail.com
Re: Problems with Oracle 11g and primary_key as rowtype IDENTITY
Posted by Miłosz Tylenda <mt...@o2.pl>.
Sebastian,
For the sequence name too long problem, you could try setting the maxAutoAssignNameLength dictionary property to the maximum sequence name length in Oracle.
You can also set the sequence name via the autoAssignSequenceName dictionary property but in this case you will rather need to create the sequence manually.
If the above fails, there is an option to override DBDictionary.getGeneratedKeySequenceName or OracleDictionary.getCreateTableSQL methods.
You might want to open a JIRA [1] issue for this problem.
Let us know if this helps.
Cheers,
Milosz
[1] https://issues.apache.org/jira/browse/OPENJPA
> Or let me re-phrase my question:
>
> Is there a possibility to define an alternative name for the sequence
> or a table when you are using the option:
> <property name="openjpa.jdbc.DBDictionary"
> value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
>
> Or is it really required at this point to rename the table's primary key ??
>
> The error message Oracle throws is like that: "could not create
> sequence because "appointmentcategory_seq" is too long name."
>
> I will also cross post this question to the Oracle support forums,
> maybe they also have an idea how to workaround that.
>
> Btw: Using
> <property name="openjpa.jdbc.SynchronizeMappings" value="buildSchema" />
>
> Also throws some error in our application using Oracle database...
> So I always need to manually delete all tables and re-run the
> application so that the tables are created (including sequences). Is
> there a way of forcing openJPA to drop all tables? In hibernate you
> would say "create" instead of "update". In openJPA I don't see any
> alternative to "buildSchema". Is there a "createSchema" ?!
> The documentation does not say so much about schema update /
> forwarding.I don't see any other option that "buildSchema" mentioned
> in the docs...
>
> Thanks
> Sebastian
>
>
>
> 2011/10/22 seba.wagner@gmail.com <se...@gmail.com>:
> > Hi,
> >
> > we are still trying to figure out what is going wrong with Oracle and openJPA.
> >
> > Our configuration now looks like that:
> > <properties>
> > <property name="openjpa.RuntimeUnenhancedClasses" value="unsupported"/>
> > <property name="openjpa.ConnectionDriverName"
> > value="oracle.jdbc.driver.OracleDriver"/>
> > <property name="openjpa.ConnectionProperties"
> > value="username=SYSTEM, password=admin"/>
> > <property name="openjpa.ConnectionURL"
> > value="jdbc:oracle:thin:@localhost:1521:openmeetings"/>
> > <property name="openjpa.jdbc.SynchronizeMappings" value="buildSchema"/>
> > <property name="openjpa.Log" value="DefaultLevel=WARN, Tool=INFO" />
> > <property name="openjpa.DataCache" value="true"/>
> > <property name="openjpa.QueryCache" value="true"/>
> > <property name="openjpa.jdbc.DBDictionary"
> > value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
> > <property name="openjpa.jdbc.QuerySQLCache" value="false"/>
> > </properties>
> >
> > UseTriggersForAutoAssign=true is the magic key to use sequences as it seems ...
> > however ...
> >
> > => The problem now is that Oracle complains that some sequences have a
> > too long column name. How can we bypass that?
> >
> >
> > Thanks
> > Sebastian
> >
> >
> >
> > 2011/10/2 seba.wagner@gmail.com <se...@gmail.com>:
> >> The guys from Oracle do (a bit offensive) argue that for such reasons
> >> sequences do exist.
> >>
> >> https://forums.oracle.com/forums/thread.jspa?messageID=9909360#9909360
> >>
> >> Quote:
> >> "Oracle has sequences for this purpose.
> >> They are either directly used in an INSERT statement, or assigned to a
> >> column by means of a trigger.."
> >>
> >> So my question might be: Why does openJPA not create sequences? Am I missing
> >> some param in my config?
> >> I have seen for MySQL there is an additional "sequence-table" created with
> >> all index. In Oracle there seems to be no such table ... while except the
> >> persistance.xml the code is 1:1 the same.
> >>
> >> Thanks,
> >> Sebastian
> >>
> >> 2011/10/1 seba.wagner@gmail.com <se...@gmail.com>
> >>>
> >>> Hi,
> >>>
> >>> when using Oracle 11g with openJPA I do get the following warning while
> >>> the tables are populated by openJPA:
> >>>
> >>> "Oracle" does not support auto-assigne d column values. The column may
> >>> not behave as desired.
> >>>
> >>> ... and the effect is obviously that, as soon as you try to add a new
> >>> column with a null value for a primary key, Oracle will throw an
> >>> Exception...
> >>>
> >>> I can read from the Issue tracker:
> >>> https://issues.apache.org/jira/browse/OPENJPA-1090
> >>>
> >>> But actually the JIRA Issue 1090 is about adding this warning, not about
> >>> how to solve it.
> >>>
> >>> What is actually the solution to fix it?
> >>> Is it database related or do we need to change our code?
> >>>
> >>> Thank you!
> >>> Sebastian
> >>>
> >>> --
> >>> Sebastian Wagner
> >>> http://www.openmeetings.de
> >>> http://www.webbase-design.de
> >>> http://www.wagner-sebastian.com
> >>> seba.wagner@gmail.com
> >>
> >>
> >>
> >> --
> >> Sebastian Wagner
> >> http://www.openmeetings.de
> >> http://www.webbase-design.de
> >> http://www.wagner-sebastian.com
> >> seba.wagner@gmail.com
> >>
> >
> >
> >
> > --
> > Sebastian Wagner
> > http://www.openmeetings.de
> > http://www.webbase-design.de
> > http://www.wagner-sebastian.com
> > seba.wagner@gmail.com
> >
>
>
>
>
Re: Problems with Oracle 11g and primary_key as rowtype IDENTITY
Posted by "seba.wagner@gmail.com" <se...@gmail.com>.
Or let me re-phrase my question:
Is there a possibility to define an alternative name for the sequence
or a table when you are using the option:
<property name="openjpa.jdbc.DBDictionary"
value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
Or is it really required at this point to rename the table's primary key ??
The error message Oracle throws is like that: "could not create
sequence because "appointmentcategory_seq" is too long name."
I will also cross post this question to the Oracle support forums,
maybe they also have an idea how to workaround that.
Btw: Using
<property name="openjpa.jdbc.SynchronizeMappings" value="buildSchema" />
Also throws some error in our application using Oracle database...
So I always need to manually delete all tables and re-run the
application so that the tables are created (including sequences). Is
there a way of forcing openJPA to drop all tables? In hibernate you
would say "create" instead of "update". In openJPA I don't see any
alternative to "buildSchema". Is there a "createSchema" ?!
The documentation does not say so much about schema update /
forwarding.I don't see any other option that "buildSchema" mentioned
in the docs...
Thanks
Sebastian
2011/10/22 seba.wagner@gmail.com <se...@gmail.com>:
> Hi,
>
> we are still trying to figure out what is going wrong with Oracle and openJPA.
>
> Our configuration now looks like that:
> <properties>
> <property name="openjpa.RuntimeUnenhancedClasses" value="unsupported"/>
> <property name="openjpa.ConnectionDriverName"
> value="oracle.jdbc.driver.OracleDriver"/>
> <property name="openjpa.ConnectionProperties"
> value="username=SYSTEM, password=admin"/>
> <property name="openjpa.ConnectionURL"
> value="jdbc:oracle:thin:@localhost:1521:openmeetings"/>
> <property name="openjpa.jdbc.SynchronizeMappings" value="buildSchema"/>
> <property name="openjpa.Log" value="DefaultLevel=WARN, Tool=INFO" />
> <property name="openjpa.DataCache" value="true"/>
> <property name="openjpa.QueryCache" value="true"/>
> <property name="openjpa.jdbc.DBDictionary"
> value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
> <property name="openjpa.jdbc.QuerySQLCache" value="false"/>
> </properties>
>
> UseTriggersForAutoAssign=true is the magic key to use sequences as it seems ...
> however ...
>
> => The problem now is that Oracle complains that some sequences have a
> too long column name. How can we bypass that?
>
>
> Thanks
> Sebastian
>
>
>
> 2011/10/2 seba.wagner@gmail.com <se...@gmail.com>:
>> The guys from Oracle do (a bit offensive) argue that for such reasons
>> sequences do exist.
>>
>> https://forums.oracle.com/forums/thread.jspa?messageID=9909360#9909360
>>
>> Quote:
>> "Oracle has sequences for this purpose.
>> They are either directly used in an INSERT statement, or assigned to a
>> column by means of a trigger.."
>>
>> So my question might be: Why does openJPA not create sequences? Am I missing
>> some param in my config?
>> I have seen for MySQL there is an additional "sequence-table" created with
>> all index. In Oracle there seems to be no such table ... while except the
>> persistance.xml the code is 1:1 the same.
>>
>> Thanks,
>> Sebastian
>>
>> 2011/10/1 seba.wagner@gmail.com <se...@gmail.com>
>>>
>>> Hi,
>>>
>>> when using Oracle 11g with openJPA I do get the following warning while
>>> the tables are populated by openJPA:
>>>
>>> "Oracle" does not support auto-assigne d column values. The column may
>>> not behave as desired.
>>>
>>> ... and the effect is obviously that, as soon as you try to add a new
>>> column with a null value for a primary key, Oracle will throw an
>>> Exception...
>>>
>>> I can read from the Issue tracker:
>>> https://issues.apache.org/jira/browse/OPENJPA-1090
>>>
>>> But actually the JIRA Issue 1090 is about adding this warning, not about
>>> how to solve it.
>>>
>>> What is actually the solution to fix it?
>>> Is it database related or do we need to change our code?
>>>
>>> Thank you!
>>> Sebastian
>>>
>>> --
>>> Sebastian Wagner
>>> http://www.openmeetings.de
>>> http://www.webbase-design.de
>>> http://www.wagner-sebastian.com
>>> seba.wagner@gmail.com
>>
>>
>>
>> --
>> Sebastian Wagner
>> http://www.openmeetings.de
>> http://www.webbase-design.de
>> http://www.wagner-sebastian.com
>> seba.wagner@gmail.com
>>
>
>
>
> --
> Sebastian Wagner
> http://www.openmeetings.de
> http://www.webbase-design.de
> http://www.wagner-sebastian.com
> seba.wagner@gmail.com
>
--
Sebastian Wagner
http://www.openmeetings.de
http://www.webbase-design.de
http://www.wagner-sebastian.com
seba.wagner@gmail.com
Re: Problems with Oracle 11g and primary_key as rowtype IDENTITY
Posted by "seba.wagner@gmail.com" <se...@gmail.com>.
Hi,
we are still trying to figure out what is going wrong with Oracle and openJPA.
Our configuration now looks like that:
<properties>
<property name="openjpa.RuntimeUnenhancedClasses" value="unsupported"/>
<property name="openjpa.ConnectionDriverName"
value="oracle.jdbc.driver.OracleDriver"/>
<property name="openjpa.ConnectionProperties"
value="username=SYSTEM, password=admin"/>
<property name="openjpa.ConnectionURL"
value="jdbc:oracle:thin:@localhost:1521:openmeetings"/>
<property name="openjpa.jdbc.SynchronizeMappings" value="buildSchema"/>
<property name="openjpa.Log" value="DefaultLevel=WARN, Tool=INFO" />
<property name="openjpa.DataCache" value="true"/>
<property name="openjpa.QueryCache" value="true"/>
<property name="openjpa.jdbc.DBDictionary"
value="oracle(UseTriggersForAutoAssign=true,BatchLimit=100)" />
<property name="openjpa.jdbc.QuerySQLCache" value="false"/>
</properties>
UseTriggersForAutoAssign=true is the magic key to use sequences as it seems ...
however ...
=> The problem now is that Oracle complains that some sequences have a
too long column name. How can we bypass that?
Thanks
Sebastian
2011/10/2 seba.wagner@gmail.com <se...@gmail.com>:
> The guys from Oracle do (a bit offensive) argue that for such reasons
> sequences do exist.
>
> https://forums.oracle.com/forums/thread.jspa?messageID=9909360#9909360
>
> Quote:
> "Oracle has sequences for this purpose.
> They are either directly used in an INSERT statement, or assigned to a
> column by means of a trigger.."
>
> So my question might be: Why does openJPA not create sequences? Am I missing
> some param in my config?
> I have seen for MySQL there is an additional "sequence-table" created with
> all index. In Oracle there seems to be no such table ... while except the
> persistance.xml the code is 1:1 the same.
>
> Thanks,
> Sebastian
>
> 2011/10/1 seba.wagner@gmail.com <se...@gmail.com>
>>
>> Hi,
>>
>> when using Oracle 11g with openJPA I do get the following warning while
>> the tables are populated by openJPA:
>>
>> "Oracle" does not support auto-assigne d column values. The column may
>> not behave as desired.
>>
>> ... and the effect is obviously that, as soon as you try to add a new
>> column with a null value for a primary key, Oracle will throw an
>> Exception...
>>
>> I can read from the Issue tracker:
>> https://issues.apache.org/jira/browse/OPENJPA-1090
>>
>> But actually the JIRA Issue 1090 is about adding this warning, not about
>> how to solve it.
>>
>> What is actually the solution to fix it?
>> Is it database related or do we need to change our code?
>>
>> Thank you!
>> Sebastian
>>
>> --
>> Sebastian Wagner
>> http://www.openmeetings.de
>> http://www.webbase-design.de
>> http://www.wagner-sebastian.com
>> seba.wagner@gmail.com
>
>
>
> --
> Sebastian Wagner
> http://www.openmeetings.de
> http://www.webbase-design.de
> http://www.wagner-sebastian.com
> seba.wagner@gmail.com
>
--
Sebastian Wagner
http://www.openmeetings.de
http://www.webbase-design.de
http://www.wagner-sebastian.com
seba.wagner@gmail.com
Re: Problems with Oracle 11g and primary_key as rowtype IDENTITY
Posted by "seba.wagner@gmail.com" <se...@gmail.com>.
The guys from Oracle do (a bit offensive) argue that for such reasons
sequences do exist.
https://forums.oracle.com/forums/thread.jspa?messageID=9909360#9909360
Quote:
"Oracle has *sequences* for this purpose.
They are either directly used in an INSERT statement, or assigned to a
column by means of a *trigger*.."
So my question might be: Why does openJPA not create sequences? Am I missing
some param in my config?
I have seen for MySQL there is an additional "sequence-table" created with
all index. In Oracle there seems to be no such table ... while except the
persistance.xml the code is 1:1 the same.
Thanks,
Sebastian
2011/10/1 seba.wagner@gmail.com <se...@gmail.com>
> Hi,
>
> when using Oracle 11g with openJPA I do get the following warning while the
> tables are populated by openJPA:
>
> "Oracle" does not support auto-assigne d column values. The column may not
> behave as desired.
>
> ... and the effect is obviously that, as soon as you try to add a new
> column with a null value for a primary key, Oracle will throw an
> Exception...
>
> I can read from the Issue tracker:
> https://issues.apache.org/jira/browse/OPENJPA-1090
>
> But actually the JIRA Issue 1090 is about adding this warning, not about
> how to solve it.
>
> What is actually the solution to fix it?
> Is it database related or do we need to change our code?
>
> Thank you!
> Sebastian
>
> --
> Sebastian Wagner
> http://www.openmeetings.de
> http://www.webbase-design.de
> http://www.wagner-sebastian.com
> seba.wagner@gmail.com
>
--
Sebastian Wagner
http://www.openmeetings.de
http://www.webbase-design.de
http://www.wagner-sebastian.com
seba.wagner@gmail.com