All language subtitles for 9. Step-by-Step DAX Measure Calculation

All right so we talked about filter context and we talked about how measures evaluate based on that

filter context.

But it's time to take a peek under the hood and try to learn about how this actual engine works.

How are these specific measure calculations happening.

So let's consider a simple matrix view like this.

In this case we're evaluating a total return's measure based on category names.

And what we want to figure out is how this particular cell accessories returns.

Number of 1115 is calculated.

So what I'm about to show you all happens behind the scenes and it happens instantly every time the

filter or context changes.

So I like to think of it as a three step process where Step one is when the filter or context is detected

and applied.

So this is where power be-I looks at the filter context for that particular value and says OK I'm being

asked to only care about rows where the category name in my product table equal accessories as a result

that product table gets filtered down to the point where you're only left with accessories Rose.

And from there we move into step two which is where that filter context category name equals accessories

flows to all downstream related tables.

So if you had a model like this where that product table or connected to data tables sales and returns

you would see both of those data tables filtered down accordingly to only the product IDs where the

category equals accessories.

So you end up with filtered down versions of both of those data tables as well.

And that brings us to step three which is one to measure formula actually evaluates against that filter

table.

So the measure that we're looking at here what it's essentially doing is it's counting the rows in that

return's data table but not the whole thing.

It's counting the rows of the filtered down version where the product category is accessories.

And as a result the number of rows that exist after the filter context has been applied is 1115.

So I'd recommend keeping this on hand just as a helpful reference in those cases where your measures

might not be calculating as you expect them to.

So that that wraps up the high level kind of conceptual part of this section of the course from here

we're going to move into daks syntax operators common function categories and then get hands on with

a bunch of different examples.