Home Page for the TeradataForum

Archives of the TeradataForum

Message Posted: Mon, 14 Feb 2005 @ 16:57:35 GMT

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

Subj:   Re: Random Amp Sampling
From:   Dieter Noeth

Anomy Anom wrote:

  I am executing explains on a much skewed table with no stats collected on it. As the parser would be doing *random* amp sampling should not I be getting different estimates for every execution or does it sample from the amp that corresponds to the table id of the table to get consistent explains? The document also says that an amp is chosen randomly.  

It's not random, it's based on the (hashed) Table ID and the result is kept in the PE's cache (AFAIK up to 4 hours). So for a given table always the same AMP is responsible to do that sampling. Even if you connect through different PEs the estimated numbers (and plans) are the same.


  <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: 15 Jun 2023