Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Wed, 06 Feb 2008 @ 16:03:11 GMT


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


Subj:   Re: Reduce loading time
 
From:   Robert D Meunier

I have to disagree with Omer on this. TPump isn't just for real time loading.

Pro's of using TPump vs. MultiLoad (not an exhaustive list):

- TPump is more versatile then MultiLoad in that you don't have to remove Join Indices prior to loading.

- TPump also doesn't count towards the number of concurrent load utilities running on Teradata so you don't have to be as concerned with exceeding the limit.

- TPump is probably faster then MultiLoad if you are touching less then 1 record per block. This has long been the Teradata mantra about TPump, but I have never done any exhaustive testing, but it's in the ballpark.


With that being said, TPump is also the most finicky of the utilities when it comes to performance. In a recent benchmark on a 4 node 5500 I was able to load 900k rows/minute. However, before I did any tuning of the TPump I was only loading 200k rows/minute. So you really have to know what you're doing if you're going to switch your MultiLoads to TPumps.

Some loading options you really have to be concerned with in TPump (once again not an exhaustive list) are:

- SESSION

- SERIALIZE (and the appropriate setting of the KEY option)

- PACK

- ARRAYSUPPORT (only available with V2R6.2 and above)


In another post there's a comment about errors. TPump's performance is seriously degraded if you have many errors going to the error table.


Robert Meunier



     
  <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