Archives of the TeradataForum
Message Posted: Mon, 17 Nov 2003 @ 19:27:46 GMT
I just learnt that in the Acquisition Phase of Fastload, the raw data is sent to AMPs, which Hashes and redistributes the records into the correct AMPs based on the Hashes.
1. On what basis, Fastload sends the raw data to AMPs...? Which data goes to which AMP, before the hashing takes place..? How the first distribution occurs ?
2. In Phase 1, records are written into the tables, as unsorted. Again, once the phase 1 is over, in phase 2, the sorting takes place and is gain written. Doesn't this cause a sort of DOUBLE I/O and create an overhead ?
Thanks and Regards,
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|