you can download the ODBC / OLEDB connector from the teradata.com site
but as far as i could see, the performance is very poor.
a query that i ran within teradata and took up to 5 minuts, in QV i aborted it after half an hour, extracting 100k lines out of 500k i was supposed to get.
but im sure im missing somthing here.... so if you find anything else keep me updated :-)
i have connected to teradata from qlikview.
first, we need to create a new datasoure in the control panel->administrative tools->data source->select the teradata driver and provide the data source name,teradata server name, username and password and save it.
then go to qlikview, in edit script, click connect and select the created data source and enter username and password, then it connects to database.
just reload it and check, it shows connected.
then you can get the tables from database from select button below the connect button in edit script window.
add the tables and reload it. you have data now from database
The direct discovery will make the whole model a whole lot slower... AFAIK it queries the DB after each selection in order to recover the possible values... isn't it more practical to query the whole thing and save it to Qvd instead?
EDIT: If you need to recover a large database then you may need to think in some sort of partial reload scenario but querying a slow database via Direct Discovery just sounds weird to me