Stacey Roach

You can get around this by creating a new project that is J2EE 1.2 instead of 1.3 and then bring in all the files from the 1.3 project. The problem is that the higher servlet engine doesn't support 4.0 data sources. You should be able to use a 5.0 data source with Oracle however.

In case anyone needs a tutorial on this, here is what I did.

I downloaded and installed the latest JDBC drivers from Teradata/NCR (version 3.01.00.08).

I edited my local hosts file (c:\\WINNT\\system32\\drivers\\etc\\hosts) and added two aliases to the server.

127.127.127.1 example1
127.127.127.1 example1cop1

I then created a JDBC provider within the Data Source tab of the .wsi screen for WebSphere. I pointed the implementation class at com.ncr.teradata.TeraConnectionPoolDataSource and added both jar files within the class path.

Then I created a JNDI 5.0 data source and used com.ibm.websphere.rsadapter.GenericDataStoreHelper as the Data Source Helper Class.

Then I added three properties, each of type java.lang.String: DSName, user, and password. DSName points to the entry in the hosts file. When I didn't do this, I kept getting an error about "default transaction mode" which I found a KB article about. Apparantly the cop1 entry is mandatory or something.

Now I had all of this setup for a few hours and still couldn't get it to work...it would fail on the executeQuery statements with a null pointer exception. This was a problem with a version of the driver (fixed in 3.01.03 (reference: DR84904)) which seemed to be a problem I was causing myself (multiple copies of different jars loaded at once while trying to debug).

Has anyone had any luck?

I need to setup a 5.0 data source with connection pooling to Teradata.

I established a Teradata JDBC provider pointing it at the com.ncr.teradata.TeraConnectionPoolDataSource class in the terajdbc4.jar file from NCR.

I then added a 5.0 data source, but I don't have or know of a Data Source Helper class for Teradata.

The options listed are:
com.ibm.websphere.rsadapter.DB2UniversalDataStoreHelper
com.ibm.websphere.rsadapter.DB2DataStoreHelper
com.ibm.websphere.rsadapter.DB2AS400DataStoreHelper
com.ibm.websphere.rsadapter.DB2390DataStoreHelper
com.ibm.websphere.rsadapter.DB2390LocalDataStoreHelper
com.ibm.websphere.rsadapter.CloudscapeDataStoreHelper
com.ibm.websphere.rsadapter.CloudscapeNetworkServerDataStoreHelper
com.ibm.websphere.rsadapter.InformixDataStoreHelper
com.ibm.websphere.rsadapter.SybaseDataStoreHelper
com.ibm.websphere.rsadapter.OracleDataStoreHelper
com.ibm.websphere.rsadapter.ConnectJDBCDataStoreHelper
com.ibm.websphere.rsadapter.WSConnectJDBCDataStoreHelper
com.ibm.websphere.rsadapter.SequeLinkDataStoreHelper
But none of those seem to work for Teradata.

I then set up the properties:
DSName, user, and password each with a type of java.lang.String

Ultimately I am getting an issue in the javax.naming.InitialContext.lookup when its trying to open the JNDI reference.
[ November 05, 2004: Message edited by: Stacey Roach ]

This sounds perfect for someone like me (just getting into learning about J2EE). Is it possible for you to explain specifically how you address the pros & cons that you present? I'm curious as to the approach and how analytical you are with it. Perhaps you can provide a basic example?

Honestly, in Java - its because I haven't found a good IDE that I like. I tried Cafe a few years back and hated it. I just loaded WSAD 5.1 on my box at work since that's what we use - hate it. I used to use JBuilder and it was alright at the time - I haven't tried it lately. I'm still using Textpad.

I did download Eclipse late last week and am finding that it is much better than whatever IBM did with it to turn it into WSAD. (Try just opening a .java file in WSAD?)

I'm typically all about IDEs if they are good ones - its just hard to find a good one. Eclipse seems to be pretty robust, so I'm starting to play with it more since I'm getting back into java with this new job (and thus how I found this site!)