Hi Benoit,

Before answering just let me propose that we keep in the loop my coleagues 
and other KIM users bu CC-ing our mails to the KIM-discussion mailing list. 
This will also make possible someone else to provide help or answer your 
questions if I am not available or cannot answer them.

> ok, no problem, he is the result i got form step 5 (./bin/toolCoreDb.cmd):
> =============================================================
> ==  COREDB - CLEARING TABLES ...
> java.sql.SQLException: ORA-00439: feature not enabled: Partitioning

So, you discovered by yourself the reason for this - the feature is added in 
Oracle10g2.

> =============================================================
> ==  COREDB - OPTIMIZING INDICES ...
> java.sql.SQLException: ORA-00439: feature not enabled: Online Index Build

The same problem - the index was not built at all so it cannot be optimized. 
The problem here is that crashing on creating that index has skipped 
creating some other indices and tables after it i.e. the schema was not 
created completely.

> I'm using ORACLE Express Edition which explains the 2 exceptions i got:
> * java.sql.SQLException: ORA-00439: feature not enabled: Partitioning
> * java.sql.SQLException: ORA-00439: feature not enabled: Online Index 
> Build

Actually we have used CORE only with Oracle10g2 Enterprise (which is free 
for development purposes) till now.

> I hope this is not an issue.

It is not, this index could be created without partitionig that will only 
make it slower. But the creation code is hardcoded in Java class. We just 
never thought of using different Oracle version or providing means of 
customizing this.

> Then, when i launch the command: /bin/toolPopulate.cmd (Step6) everything
> went fine.

Which is strange because I would expect it to crash too at the final step 
when indices are being optimized again.

> But now i'm blocked at another point.
> When i try to run the example: CoreSearchAndTimelines.java
> i got the following output:

As I said - the schema was not created completely. It is a question of 
tracing the code to see exactly which tables and indices are created after 
the one that crashed, but anyway, ommiting any of them will lead to further 
problems at some point be it adding documents or querying them.

Well, I can think of two options to solve this:
(1) the best one, of course, is to install Oracle10g2 as this will save you 
futher problems (I mean when we workaround this something new may arise just 
because we haven't tested CORE on another versions)
(2) and the workaround is to patch the code that creates the index - I could 
catch an exception on creating exactly this index with partitions and as an 
alternative try to create a one without - and I can send you the recompiled 
KIM.jar when finished (this may take some time as I will have to revert to 
and patch the version that you're probably using  1.7.12.15)

I guess you will prefer (2) but I would still like to hear what do you 
think.
Regards,
Ilian 


_______________________________________________
NOTE: Please REPLY TO ALL to ensure that your reply reaches all members of this 
mailing list.

KIM-discussion mailing list
[email protected]
http://ontotext.com/mailman/listinfo/kim-discussion_ontotext.com

Reply via email to