I'm currently using ODI 11g to import into Oracle, via CSV files, records from Mainframe Adabas table views. This is being done successfully.
The point is that I'm trying now to send back to a mainframe application via CSV the records that, for a reason or other, could not be imported into Oracle and are stored in the ETL's error tables.
I'm trying to use the same process, in this case backwards, to export the data from the error tables to a CSV file, which is to be imported by the mainframe application into Adabas.
I successfully imported via reverse engineering the structure of the error table to be my source base. I've set up new physical e and logical models to be used by this process. I've also created the interface.
My problem is that when I try to save the interface, it gives me a fatal error saying that I don't have an "LKM selected for this origin set".
When I try to set the LKM in Flow tab, it doesn't give me any option at LKM Selector.
I'm quite green on ODI and have no idea how to solve this problem, so any insights would be most appreciated.
Thanks all!
You need to change the location where transformations will occur. Currently the interface is trying to move all data to the file technology and process it there. But it's easier to work the other way around and make the database do the job. To do so, go on the overview pane of your interface and select "Staging Area Different From Target" checkbox, then select the logical schema of your Oracle source below.
On the Flow tab, click on your target and select the following IKM : "IKM SQL to File Append". This is a multi-technology IKM which means you won't need an LKM anymore to move data from source to target.