Archives of the TeradataForum

Message Posted: Sat, 30 Dec 2006 @ 11:41:43 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

It's not easy. BTEQ is not aware of the operating system. It can perform an OS command, but that command cannot reference BTEQ variables, so you can't, for instance, echo records inserted into a file for further processing (as you can do with MLOAD).

We have implemented a log "scraping" utility in unix. When we run a BTEQ command that we want to trap, we put a standard tag name ahead of, and following the statement. We also run a 'select current_timestamp(0);' statement before and after the statement. Then, the BTEQ output is routed to a temporary logfile where it is scanned looking for the beginning tag, beginning timestamp, BTEQ standard output (record counts) and ending timestamp. We gather the information and write it to a log file that we later store in table form.

Hope this is helpful. For what it is worth, the Teradata Parallel Transporter falls more in line with BTEQ than MultiLoad. I raised this as an issue at this years Partners conference.

JT

	<Prev	Next>		<<First	<Prev	Next>	Last>>