CSV Import Error

Is there a trick to getting TDP to import a csv file it created itself into a table that it created itself? I exported data to csv from Oracle and tried to import it into Teradata and it continues to fail. I opened the file in a text editor and the format seems to check out. I used the import wizard to import it to Teradata. The delimiter and column mappings are correct yet I keep getting errors indicating row field in prescan step. Is there a trick to making TDP happy with its own files?

Hard to tell from your description whether this is a TDP issue. Are you certain that the datatypes from the import file are compatible with Teradata? It's very likely that some Oracle datatypes do not map well to any of the Teradata types.

Also, assuming that you're using the Professional edition of TDP, there is an Import/Export Wizard that attempts to do cross-migration in one shot... again, assuming datatypes that play well from source to target.

You may have to utilize TDP's Transform & Cleanse engine (or equivalent capabilities if you have other products that do this) to groom some of the data types on the import data file to map better with the target data.

The problems I've had with data type conversion I've been able to resolve with unicode. It is a professional edition version 5.0.7.106. In this scenario TDP exported from Oracle to a csv file. When I attempted to import it into Teradata it failed, but the data itself was primarily varchar with a few int and date values.

Would help if you shared any error messages you get, or information panels that pop up when the snag happens. Also, upgrade to 5.1, and check out the Export/Import wizard. The datatypes you mention seem common enough to avoid problems (although I suspect the date field types, if anything). If issue persists, I would suggest sending to our Support team, who may be able to reproduce the issue. I do see in our Knowledge base that others have had issues with Teradata and CSV file imports, so Support may have some packaged guidance right away.

Thanks, I'll work on gathering some examples.