User is running a query that creates a large resultset (145 million rows). The query itself completes in about 2 minutes, but sending the 145 million row resultset across a wide network is taking a VERY long time. Is there a way to improve the performance of a query's RESPONDING phase (i.e. sending the results back to the client)? I worked with Oracle (sorry) in a previous life and it had a feature called arraysize which directed Oracle to put multiple result rows into an array before sending the array across the network to the client. Arraysize improved performance TREMENDOUSLY. Does Teradata have anything like arraysize, or anything else that improves the performance of a query's RESPONDING phase?
As noted, your solution for this is TPT Export/FastExport.
The real cause of the problem you're facing is using tools like Bteq, SQLA or Studio the export is done on a single session. TPT/Fastexport uses multiple sessions.
In any of the above tools, the data is 'blocked' so that each network message typically contains multiple rows.
FYI: Teradata has 'array list' support but only for loading, not for exporting data.