Tag Archives: Performance Tuning

A Method for Handling Process Skew with Unknown Defaults

Back in the post Distributed Joins, Process Skew I mentioned a common cause of process skew were default values such as -1 for an “Unknown” customer creating skew when a redistribution happened on that customer key and there was a significant … Continue reading

Posted in Performance Tuning for Netezza | Tagged , , | Leave a comment

Distribution, Not Just for Joins

So I think in previous posts the need for distribution and therefore the benefit for colocation of data to handle joins has been pretty well covered.  But there are several other scenarios where data needs to be distributed to process. … Continue reading

Posted in Performance Tuning for Netezza | Tagged , , , , | 4 Comments

Distributed Joins, Process Skew

This post will build on concepts introduced in the Distributed Joins, The Basics and Distributed Joins, Modeling for Colocation posts.  I’m assuming some familiarity with table skew, where a distribution key is chosen where some key values has significantly more rows than … Continue reading

Posted in Best Practices, Performance Tuning for Netezza | Tagged , , , , | 8 Comments

Distributed Joins, Modeling for Colocation

This post will build on my previous post Distributed Joins, The Basics.   So we will extend the example from a customer table and an orders table, to a customer table , orders table, and order_items table.  Logically the customer skey  … Continue reading

Posted in Best Practices, Performance Tuning for Netezza | Tagged , , , , | 3 Comments

Distributed Joins, The Basics

This post is really about some base understanding that will be needed to understand some later performance methods.  I find to understand database performance it is helpful to understand and think how things are done at the most basic level.  … Continue reading

Posted in Best Practices, Performance Tuning for Netezza | Tagged , , , , , , | Leave a comment