Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Tue, 15 Jun 2004 @ 16:57:49 GMT


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


Subj:   Re: Mload delete and Bteq Delete
 
From:   Victor Sokovin

Uli,

  The main difference (in simple words) is that the Multiload Delete does not write into the journal tables and is therefore faster. You are not allowed to use PI delete statements in a Multiload Delete (where you have to qualify PI in a normal Multiload calling a delete statement). If you need to delete huge volumes from a table (but not all rows). Multiload Delete is quite good to do so.  



I agree that ML is at its best in the cases you have described and it would not even occur to me to use ML to delete _all_ rows from a table but while we are at this topic it might be useful to understand what will happen if, say, a code generator submits such a DELETE statement to ML. I see three possibilities which I don't have time to try out right now so I will just write what I think will happen:

1. DELETE is specified before the BEGIN MLOAD command (non-"fast path" mode).

ML will accept the statement. Although it would normally use the Transient Journal in such cases, the DELETE would be fast because TD understands that all the rows will have to be deleted.

2. DELETE is specified in BEGIN IMPORT MLOAD.

The statement will fail because TD expects to see the full PI specification in the statement etc.

3. DELETE is specified in BEGIN DELETE MLOAD. It will be accepted and it will be fast.


As I said, this is just what I think will happen. It would be good to have this confirmed.


Regards,

Victor



     
  <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: 23 Jun 2019