We have a custom built ETL application which loads more than 30,000 files per day in near real time with number of rows in each file ranging from 100 to 89000 using Bteq imports with optimal pack factors and single sessions. They run like a charm.
But we cant use them as Fastload works. As I read it some where Fastload is the darling of Teradata's all utilities and systems. Because in another near real time system where we load 70 files per hour with rows ranging from 1,00,000 rows to 9,00,000 rows per file, the bteq import takes huge amount of time and shoots out of that hour to load. Where as when you use fastload it loads in 6-10 minutes for all 70 files. So tool for purpose.
in TPT stream operator, i am getting a deadlock on SET UPI table, i have tried SERIALIZE on (UPI) , but still i am getting the message in qrylog table "Transaction ABORTed due to deadlock. " , but the job completed sucessfully.