I was thinking specifically of trying something of the sort
"INSERT /*+ append nologging */ INTO email_demographic
(SELECT /*+ parallel( a, 4 ) */ *
FROM TABLE( CAST( etl.transform( CURSOR(SELECT * FROM ext_tab ))
AS email_demo_nt_t)) a);"
, as described on the second-to-last page (I agree that nologging is
not a valid hint, he probably meant to make the email_demographic
nologging, with "alter table ..."). But besides that, the paper seemed
to include nice advice on large volume import of data (but I admit I've
just read the paper, and not tried anything myself).
Daniel
Received on Mon Jun 06 2005 - 15:42:49 CDT
