Quoting from my popular post on performance gotchas: You probably want to try to find a split that gives roughly the same number of unique  values in each column, so don’t constrain your self to just “dollars and cents” or even first 5 digits, last 5 digits.  This one… you have to test! The goal here is better compression by having less distinct values. What I was talking about is splitting a single column of dollars and cents, into separate columns that would each have less distinct values.  Continue Reading

As referenced in my previous post, the numbers I was getting for memory usage from Kasper’s VBA Script did not make sense to me.  After taking the time to better understand how Power Pivot stores data… it became clear to me why. Let’s look at what the script does, and does not, include… and figure out how much we care. Continue Reading

Intro When I start writing a blog article, I don’t exactly know what to expect.  Though… this is basically true in everyday conversation with me, as well.  It’s hard to know what words are about to fly out of my mouth.  It’s scary for everyone involved.   That said… I expect this article to be kind of… epic.  Equal parts technical, totally made up, and totally useful As mentioned in this previous post on performance gotchas, … well, many things… but mostly that Power Pivot performance issues Continue Reading

One of my more common consults is for “my workbook is slow”.  Occasionally, I will have to dig into the actual measures, but more typically, I just check for a few common issues that can have a large negative impact on performance. Before we dig into those, let’s just take one moment to state the obvious.  Power Pivot is an in-memory database.  Almost by definition… if you want things to run faster… use less memory!   When I am looking at the customers model, the overriding question I am asking Continue Reading