PowerPivot – the ULTIMATE anti-bloat feature

Howdy again, folks. Jeff Weir here, borrowing the keys to the blog off Chandoo again. (Hopefully I don’t scratch it again).

How remiss of me…jacob reminded me in the comments of my previous BLOATED post on good spreadsheet anti-bloat practices that I completely missed one of Excel’s newest and most efficient anti-bloat features: PowerPivot. So today’s post is going to rectify that. In less words, I promise.

Does my data look flat in this?

In Rob Collie’s excellent book DAX Formulas for PowerPivot – the Excel Pro’s Guide to Mastering DAX, Rob makes the point that before PowerPivot came along, Excel pros spent lots of their time ‘flattening’ data in order to feed their pivots. In fact, many Excel Pro’s had become Excel Pro’s largely on the back of those data flattening skills.

What does he mean by ‘flattening’? Well, PivotTables are such finicky eaters that they only like digital Pizza. That is, if you want a PivotTable to fully digest your data directly from the worksheet, then you need to lay that data out in a hierarchical structure that obfuscating geeks like to call a flat file. (You or I call a flat file a table. That’s why we’re not geeks.)

A PivotTable’s rather restrictive diet reminds me of this joke:
Question: What do you feed someone with ebola, SARS, and Swine flu?
Answer: Anything that fits under the door.

So your picky PivotTable will only eat flat, boring old Tables. In fact, it will only eat ONE table, and that table better have good labeling of all the ingredients (i.e. column headers) or your precious PivotTable will not even open it’s mouth.

Which is a problem, because the BOSS just ordered you to serve up some crazy concoction that isn’t even on your regular menu. The BOSS wants you to mix a little bit of this table with a tiny bit of that table, then add a sprinkling of some other table over the top as garnish. And the BOSS expects you to slam all this into your pre-heated PivotOven for a quick bake at 2.30GHz for no more than a few minutes, and then serve it up to the BOSS right away. Because the BOSS is hungry for data, and the BOSS is hungry NOW, DAMMIT!

So what did you do? You used as many VLOOKUPS as you have rows in your final flat data-set to join just one column of one of those additional tables onto the first table, didn’t you. And then you repeated this VLOOKUP frenzy for each and every other column that you ended up bringing into your steam-rolled mega-flat pivot-ready data-set. All of which resulted in one very bloated filesize, compared to the original footprint of the underlying tables.

And while you managed to serve up the order just in time, boy did you make a mess back in the kitchen. Formulas everywhere, and the whole joint is slowing down as a result. What’s worse, the BOSS liked the taste of what you just served up. So you’ll be working in the same messy kitchen next week to refresh it, unless you tidy up somehow.

Let’s face it…it’s such a complete mess, that you’re screwed.

Or are you?

PowerPivot….No fast data joint should be without it!

If PowerPivot was marketed on the Shopping Channel, then some obnoxiously loud voice would say something like this about it:
It slices. It dices. It joins. But wait, there’s more!
It cooks. It cleans. It washes up. It takes up practically no bench-space. But wait, there’s STILL more!

In fact there’s so much more, that that’s a subject for another post. Fortunately Chandoo already wrote it: What is Power Pivot – an Introduction. (Chandoo, that title is way too descriptive. You’ll never make Class 1 Geek unless you learn to obfuscate, my man).

Give that link a spin, because this product lives up to it’s hype. Indeed, no modern fast data joint should be without it. Emphasis on modern though, because you’ll need Excel 2010 or later in order to use this bloat-busting add-in.

But back to how it helps with bloat, the subject of this post. PowerPivot cuts through potential bloat, because it is a lot less fussy than Old-School-Pivots about what it eats:

• It allows you to create pivots on the fly from any mix of multiple data sources – Access, SQL, Excel Tables, Web Data, etc – and then effortlessly slice, dice, and navigate to your hearts content.
• You can incorporate/mash up additional data sources at any point.
• You can create very powerful calculated fields within PivotTables that simply are not possible to replicate with in traditional pivots.
• All without ship-loads of VLOOKUPS.

In fact, Rob Collie – master of both PowerPivot and understatement – has a great video showing how PowerPivot is the answer to “the dreaded VLOOKUP problem, among other things” in his post Be Gone, Scary VLOOKUP”.

So it does away with all those nasty VLOOKUPs. But that’s not the half of it…PowerPivot has some amazing data compression stuff going on under the hood too! (Check out Rob’s post Surprising Example of PowerPivot Compression for more on this.)

Okay, I’m convinced. But I’m a little scared, too…

If you want help to learn PowerPivot, then help is at hand: Check out Chandoo’s Advanced Excel & Power Pivot Training Classes. Rob Collie puts in a guest appearance in one of the modules, too. (And I think that you get a copy of his great book as part of the course fee.)

But before I return you to your regular schedule, be warned: Chandoo has the following public service message on his PowerPivot landing page that you might want to consider, if your boss is attractive as mine is:
Warning: Learning Excel and Powerpivot might suddenly make you boss fall in love with you.

Indeed, that is a good warning that I will heed, Chandoo. I’m burning Rob’s PowerPivot books as you read this.

About the Author.

Jeff Weir – a local of Galactic North up there in Windy Wellington, New Zealand – is more volatile than INDIRECT and more random than RAND. In fact, his state of mind can be pretty much summed up by this:

=NOT(EVEN(PROPER(OR(RIGHT(TODAY())))))

That’s right, pure #VALUE!

Find out more at http:www.heavydutydecisions.co.nz