SQL Loder problem

tsamui · Post by **tsamui** » Tue Dec 11, 2007 10:55 am

Hi All,
In my job I am using db2 as a source and oracle as the target. The job truncates the target table first and then loads the data. To run this job I am getting some sqlldr error. I am giving some log entry from datastage log.

Message: stg_CUSTOMERS: When checking operator: The -index rebuild option has been included;
since the user has not set the environment variable APT_ORACLE_LOAD_OPTIONS,
Orchestrate by default sets the options DIRECT and PARALLEL to TRUE, and
option SKIP_INDEX_MAINTENANCE to YES.

Message: stg_CUSTOMERS,3: SQL*Loader-926: OCI error while uldlfca:OCIDirPathColArrayLoadStream for table MDSTAGE.CUSTOMERS
Message: stg_CUSTOMERS,3: SQL*Loader-2026: the load was aborted because SQL Loader cannot continue.
Message: stg_CUSTOMERS,0: SQL*Loader-925: Error while uldlgs: OCIStmtExecute (ptc_hp)
Message: stg_CUSTOMERS,0: ORA-03114: not connected to ORACLE

I am getting this kind of sqlldr error.
Note: that job ran successfully multiple times before.

Please suggest what is the problem? And how to solve it?

ray.wurlod · Post by **ray.wurlod** » Wed Dec 12, 2007 1:30 am

What part of "not connected to ORACLE" was unclear?

chulett · Post by **chulett** » Thu Dec 13, 2007 7:20 am

The 'not connected' part is a result of the other errors. Suggest you run this by your Oracle DBA and see what they have to say about it.

tsamui · Post by **tsamui** » Sat Dec 15, 2007 5:50 am

You are right chullet. The DBA also told the same thing.
Actually there were 18 million records in the source and we are using complex join also. When we decrease the no of records, the job is running fine. The environment is test environment and I think this environment is not capable enough to handle the huge records.
Any comment?

tsamui · Post by **tsamui** » Sat Dec 15, 2007 5:51 am

You are right chullet. The DBA also told the same thing.
Actually there were 18 million records in the source and we are using complex join also. When we decrease the no of records, the job is running fine. The environment is test environment and I think this environment is not capable enough to handle the huge records.
Any comment?

chulett · Post by **chulett** » Sat Dec 15, 2007 8:07 am

Your DBA should be able to handle configuring the instance to handle the load, whatever that entails. More details may be in the sqlldr log file.