Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Wed, 14 Mar 2001 @ 15:29:17 GMT


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


Subj:   Re: Secondary Indexes
 
From:   David Wellman

Howard,

Looks like we noth made an oops !

The calculation should have been;

- number of rows per week-number (i.e. number of rows per NUSI data value) = A

- number of AMPs = B

- 48kb / 8 = 6144 = C (the 8000 I used earlier was the based on a full-size 64kb block, but typically your nusi blocks won't be full size)

==> number of NUSI data blocks per data value @= (A / B) / C = (600000 / 32) / 6144 = 3

This is the number of data blocks that you'll add to a NUSI for every week number that you hold. By itself, I wouldn't consider this to be a prohibitive number of data blocks per NUSI data value.

But, coming back to your original question ... Based on the sites I've worked with, I've no problem with 3 or 4 nusi's like this on a table. I'd be carefull about adding too many more (even though they're fairly 'light' nusi's). Maybe you have to fall back on the old method of adding one and trying it. If the maintenance performance with one more is acceptable, then try adding a second etc until your maintenance performance is unacceptable. Naturally, between adding nusi's you'll also be checking your query performance to ensure that these are improving appropriately...


Cheers,

Dave



     
  <Prev Next>   <<First <Prev Next> Last>>  
 
 
 
 
 
 
 
 
  
  Top Home Privacy Feedback  
 
 
Copyright for the TeradataForum (TDATA-L), Manta BlueSky    
Copyright 2016 - All Rights Reserved    
Last Modified: 27 Dec 2016