“Connection Pool” window doesn’t appear during import

www.obieefans.com 


Q: I’m doing the “create repository” tutorial and when I press the “Import” button the “Connection Pool” window doesn’t appear.  Have I done something wrong?

No you haven’t.  There are two different methods of importing schema metadata.  What the tutorial describes is a “hot-potch” of both, so it’s no wonder you’re confused.

Method 1.  Let’s assume you’ve followed the tutorial instructions to create an “ODBC Data Source” called “SH” that contains connect information for the “SH” schema using a “tnsnames.ora” entry of “ORCL”.  Continuing to follow the tutorial, when you select “Import => from Database” you see “ODBC 3.5” as the default “Connection Type”.  You then select “SH” as the DSN.  What you are doing here is identifying the “ODBC Data Source” that you defined earlier.  When you press “OK” to bring up the “Import” window and select your tables you already have all the information needed to perform the import.  So when you press the “Import” button the import starts and completes without any “Connection Pool” window appearing.  You’ll see the name you gave to the “ODBC Data Source”, “SH”, appear as a node in the “Physical Layer”.  If you expand this node you’ll see that the tables have already been imported (the tutorial is incorrect in saying you have to do something else for the import to proceed).  Just press the “Close” button in the “Import” window and you’re done.  Now right-click on the “Connection Pool” under the “SH” node and select “Properties”.  The “Call Interface” will be set to “ODBC 3.5” (the tutorial is wrong here) and the “Data Source Name” will be set to “SH” or whatever name you used when defining the “ODBC Data Source”.  You don’t actually need to change these values: they worked successfully in allowing you to import the metadata, and they therefore provide an indirect means of connecting to the Oracle database – via the “ODBC Data Source”.  However, changing the “Call Interface” to “OCI 10g/11g” and the “Data Source Name” to the appropriate “tnsnames.ora” entry for your database, say “ORCL”, makes sense as it will effectively bypass the “ODBC Data Source” when you connect to the Oracle database in the future.

Method 2.  But, defining an “ODBC Data Source” is redundant.  The statement in the tutorial “An ODBC data source is needed to import schema information into an Oracle BI repository” is incorrect.  All you have to do to import data is to navigate to “Import => from Database” as before.  Then instead of accepting the default “Connection Type” of “ODBC 3.5”, select a “Connection Type” of “OCI 10g/11g” from the LOV.  Then enter a “TNS Name” of “ORCL”, a “User Name” of “SH”, and the appropriate password.  Select your tables from the “Import” pop-up that appears and press “Import”.  In this case the “Connection Pool” window will appear (this is where the tutorial is getting confused).  Press “OK” and the import will start.  Press “Close”.  The only difference is that the node in the “Physical Layer” will be named “ORCL” instead of “SH”.