Archives of the TeradataForum
Message Posted: Mon, 10 Dec 2001 @ 13:52:06 GMT
we did this 2 weeks ago, moving data from Oracle on an HP box to Teradata. The mechanism was 1) create named pipe, 2) start fastload sucking on pipe in background, 3) using sqlplus, write to a spool file, which is the named The whole bunch can run in the background, so multiple jobs are theoretically possible.
In our case both Oracle and fastload were on the same box. If fastload is on TD only, then RCP would be an option, although I'm not sure if FL can read the file on the remote while it's being written.
Writing to spool is painfully slow; for larger data volumes (we only moved 20GB), a Pro-C program might be better. It is imperative the spool output and the fastload layout match 100% (multiload is apparently a bit more forgiving). Since getting data out of Oracle is the bottleneck, going via the mainframe really won't speed things up.
Any other suggestions available would be very welcome, since this topic is sure to rear its head again.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|