ORA-014161 - Help with Bug on Oracle Driver

[clebert.suconic]

I'm testing JBossMessaging for the release process, and I found what seems to be a bug on Oracle driver:


"java.sql.SQLException: ORA-01461: can bind a LONG value only for insert into a LONG column"


I have a simple statement, insterting values into JBM_MSG:


INSERT INTO JBM_MSG (MESSAGE_ID, RELIABLE, EXPIRATION, TIMESTAMP, PRIORITY, HEADERS, PAYLOAD, TYPE) SELECT ?, ?, ?, ?, ?, ?, ?, ? FROM JBM_DUAL WHERE NOT EXISTS (SELECT MESSAGE_ID FROM JBM_MSG WHERE MESSAGE_ID = ?)


And this can't work when using LONGs on queries.


This is clearly a bug on the driver :-( Anyone know how to fix that?


I have googled for this of course, and the only conclusion I found was.. "this is a bug on the driver" and nothing more besides that.

It seems that everybody fixed it by just avoiding it the long.. replacing the long usage by store procedures or something else.

[apwalker]

What version of Oracle and what version of the driver are you using. I had a similar problem with Oracel 9i and the problem was fixed by using the 10g driver which is backwards compatible. So you might want to try the latest driver for 11g

[clebert.suconic]

I'm not sure what is the version (I will find out about that tomorrow during business hours... I'm using a node from our QA department)...

The problem is related to the Insert with Select.. and the Blob is on the Select clause... It looks like Oracle don't support binding blob values on a Select? (which is kind of weird though).


For now (I mean... for CR2).. I'm ignoring duplicate keys on the storeMessage method.. I will look for a way to change it before GA.


Thanks a lot...


Clebert

Ps: I will update the thread when I have discovered the versions...

[clebert.suconic]

Just for the record.. I'm really disappointed with the Oracle driver.

This is a very basic feature IMO, I have found lots of references and complains on the internet about this and they don't seem to care about it.

I will just avoid binding BLOBS on Select clauses on Oracle, as I don't want to deal with support cases on this later. The consequence will be an "if Oracle" on the code.

[clebert.suconic]

I would be able to simply use insert and ignore for duplicate keys.. but Postgres would invalidate the transaction on those cases, throwing this message:


"25P02, current transaction is aborted, commands ignored"

[clebert.suconic]

Update: I've create a JIRA for this:

http://jira.jboss.com/jira/browse/JBMESSAGING-1060