Chris Webb's BI Blog

Analysis Services, MDX, PowerPivot, DAX and anything BI-related

Archive for the ‘Analysis Services’ Category

MDX Solve Order, SCOPE_ISOLATION and the Aggregate() function

with one comment

Solve order in MDX is a mess. Back in the good old days of Analysis Services 2000 it was a difficult concept but at least comprehensible; unfortunately when Analysis Services 2005 was released a well-intentioned attempt at making it easier to work with in fact ended up making things much, much worse. In this post I’m going to summarise everything I know about solve order in MDX to try to make this complicated topic a little bit easier to understand.

If you’re an experienced MDXer, at this point you’ll probably lose interest because you think you know everything there is to know about solve order already. Up until two weeks ago that’s what I though too, so even if you know everything I say in the first half of this post keep reading – there’s some new stuff at the end I’ve only just found out about.

Let’s start with a super-simple cube built from a single table, with two measures (Sales Amount and Cost Amount) and a Product dimension containing a single attribute hierarchy with two members (Apples and Oranges). Everything is built from the following table:

image

Solve Order and calculated members in the WITH clause

To understand what solve order is and how it can be manipulated, let’s start off looking at an example that uses only calculated members in the WITH clause of a query. Consider the following:

WITH

 

MEMBER [Measures].[Cost %] AS

DIVIDE([Measures].[Cost Amount],[Measures].[Sales Amount]),

FORMAT_STRING='0.0%'

 

MEMBER [Product].[Product].[Total Fruit] AS

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]})

 

SELECT

{[Measures].[Sales Amount],

[Measures].[Cost Amount],

MEASURES.[Cost %]}

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[Total Fruit]}

ON ROWS

FROM SALES

There are two calculated members here:

  • Cost % divides Cost Amount by Sales Amount to find the percentage that costs make up of the sales amount
  • Total Fruit sums up the values for Apples and Oranges

The output of the query is as follows:

image

Solve order controls the order that MDX calculations are evaluated when two or more of them overlap in the same cell. In this case Cost % and Total Fruit are both evaluated in the bottom right-hand cell; Total Fruit is calculated first, giving the values of 30 for Sales Amount and 21 for Cost Amount, and Cost % is calculated after that. The bottom right-hand cell is the only cell where these two calculations overlap and the only cell where solve order is relevant in this query.

In this case, 70% is the value you would expect to get. You, however, can control solve order for calculations in the WITH clause by setting the SOLVE_ORDER property for each calculated member, like so:

WITH

 

MEMBER [Measures].[Cost %] AS

DIVIDE([Measures].[Cost Amount],[Measures].[Sales Amount]),

FORMAT_STRING='0.0%',

SOLVE_ORDER=1

 

MEMBER [Product].[Product].[Total Fruit] AS

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]}),

SOLVE_ORDER=2

 

SELECT

{[Measures].[Sales Amount],

[Measures].[Cost Amount],

MEASURES.[Cost %]}

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[Total Fruit]}

ON ROWS

FROM SALES

image

Now the value in the bottom right-hand corner is 135% instead of 70%: Cost % is calculated first, then Total Fruit second so 60%+75%=135%. The SOLVE_ORDER property of a calculated member is an integer value, and the lower the SOLVE_ORDER value the earlier the calculation will be evaluated, so with Cost % having a solve order of 1 and Total Fruit having a solve order of 2, this forces Cost % to be calculated first now even though in this case it gives what is clearly an ‘incorrect’ result.

Solve Order and calculated members defined on the cube

Things now get a bit more complicated. There’s a different way of controlling solve order if your calculations are defined on the cube itself: in this case, solve order is determined by the order that the calculations appear on the Calculations tab. So if the calculations tab of the Cube Editor contains the calculations in this order:

CREATE MEMBER CURRENTCUBE.[Measures].[Cost %] AS

DIVIDE([Measures].[Cost Amount],[Measures].[Sales Amount]),

FORMAT_STRING='0.0%';

 

CREATE MEMBER CURRENTCUBE.[Product].[Product].[Total Fruit] AS

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]});

image

…and you run the following query:

SELECT

{[Measures].[Sales Amount],

[Measures].[Cost Amount],

MEASURES.[Cost %]}

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[Total Fruit]}

ON ROWS

FROM SALES

You get the incorrect result again:

image

…but if you change the order of the calculations so that Total Fruit comes first:…and rerun the same query, you get the correct results:

image

image

The SOLVE_ORDER property can also be used with calculations defined on the cube to override the effect of the order of calculations. So defining the following calculations on the cube:

CREATE MEMBER CURRENTCUBE.MEASURES.[Cost %] AS

DIVIDE([Measures].[Cost Amount], [Measures].[Sales Amount]),

FORMAT_STRING='PERCENT', SOLVE_ORDER=2;

 

CREATE MEMBER CURRENTCUBE.[Product].[Product].[Total Fruit] AS

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]}), SOLVE_ORDER=1;

image

…means that, even though Total Fruit comes after Cost % on the Calculations tab, because it has a lower solve order set using the SOLVE_ORDER property it is evaluated before Cost % and the query still returns the correct value:

image

Solve order and calculations defined in the WITH clause and on the cube

What happens if some calculations are defined on the cube, and some are defined in the WITH clause of a query? By default, calculations defined on the cube always have a lower solve order than calculations defined in the WITH clause of a query; the SOLVE_ORDER property has no effect here. So if Total Fruit is defined in the WITH clause and Cost % on the cube, you get the incorrect result:

image

WITH

 

MEMBER [Product].[Product].[Total Fruit] AS

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]})

 

SELECT

{[Measures].[Sales Amount],

[Measures].[Cost Amount],

MEASURES.[Cost %]}

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[Total Fruit]}

ON ROWS

FROM SALES

 

image

Of course, if Total Fruit is defined on the cube and Cost % is defined in the WITH clause you will get the correct answer. However, usually measures like Cost % are defined on the cube and it’s calculations like Total Fruit, which define custom groupings, that are defined on an ad hoc basis in the WITH clause. This is a problem.

The SCOPE_ISOLATION property

This default behaviour of calculations defined on the cube always having a lower solve order than calculations in the WITH clause can be overridden using the SCOPE_ISOLATION property. Setting SCOPE_ISOLATION=CUBE for a calculated member defined in the WITH clause will give that calculated member a lower solve order than any calculations defined on the cube. So, with Cost % still defined on the cube the following query now gives the correct results:

WITH

 

MEMBER [Product].[Product].[Total Fruit] AS

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]}),

SCOPE_ISOLATION=CUBE

 

SELECT

{[Measures].[Sales Amount],

[Measures].[Cost Amount],

MEASURES.[Cost %]}

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[Total Fruit]}

ON ROWS

FROM SALES

image

The Aggregate() function

Using the MDX Aggregate() function (and in fact also the VisualTotals() function – but you probably won’t ever want to use it) inside a calculation has a similar effect to the SCOPE_ISOLATION property in that it forces a calculation to be evaluated at a lower solve order than anything else. Therefore, in the previous example, instead of using the SCOPE_ISOLATION property you can change the calculation to use the Aggregate() function instead of Sum() and get the correct results:

WITH

 

MEMBER [Product].[Product].[Total Fruit] AS

AGGREGATE({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]})

 

SELECT

{[Measures].[Sales Amount],

[Measures].[Cost Amount],

MEASURES.[Cost %]}

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[Total Fruit]}

ON ROWS

FROM SALES

image

The general rule is, therefore, whenever you are creating custom-grouping type calculated members like Total Fruit in the WITH clause of a query, to use the Aggregate() function rather than Sum(). The fact that Aggregate() takes into account the AggregateFunction property of each measure on the cube (so that distinct count, min and max measures are dealt with correctly) is another good reason to use it.

Using the Aggregate() function in calculations defined on the cube has the same effect. Even when the Total Fruit calculated member is defined after Cost % on the Calculations tab, as here:

image

…so long as Total Fruit uses the Aggregate() function, running the test query gives the correct result:

SELECT

{[Measures].[Sales Amount],

[Measures].[Cost Amount],

MEASURES.[Cost %]}

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[Total Fruit]}

ON ROWS

FROM SALES

image

 

There are some very interesting details about the way Aggregate() changes solve order though.

First of all, using the Aggregate() function in a calculated member doesn’t change the solve order of the whole calculation, just the part of the calculation that uses the Aggregate() function. With the following calculations defined on the cube:

CREATE MEMBER CURRENTCUBE.[Measures].[Cost %] AS

DIVIDE([Measures].[Cost Amount],[Measures].[Sales Amount]),

FORMAT_STRING='0.0%';

 

CREATE MEMBER CURRENTCUBE.[Product].[Product].[One Aggregate] AS

AGGREGATE({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]});

 

CREATE MEMBER CURRENTCUBE.[Product].[Product].[One Sum] AS

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]});

 

CREATE MEMBER CURRENTCUBE.[Product].[Product].[Two Aggregates] AS

AGGREGATE({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]})

+

AGGREGATE({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]});

 

CREATE MEMBER CURRENTCUBE.[Product].[Product].[Two Sums] AS

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]})

+

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]});

 

CREATE MEMBER CURRENTCUBE.[Product].[Product].[One Aggregate One Sum] AS

AGGREGATE({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]})

+

SUM({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]});

…running the following query:

SELECT

{[Measures].[Sales Amount],

[Measures].[Cost Amount],

MEASURES.[Cost %]}

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[One Aggregate],

[Product].[Product].[One Sum],

[Product].[Product].[Two Aggregates],

[Product].[Product].[Two Sums],

[Product].[Product].[One Aggregate One Sum]}

ON ROWS

FROM SALES

…gives these results:

image

The value returned for the calculation [One Aggregate One Sum], which contains an Aggregate() and a Sum(), shows that the value returned by the Aggregate() is evaluated at a different solve order than the value returned by Sum(), even if they are inside the same calculated member.

Furthermore, in some very obscure cases the contents of the set passed to the Aggregate() function determine whether its special solve order behaviour happens or not. I don’t know for sure what all those cases are but I have seen this happen with time utility (aka date tool aka shell) dimensions. Here’s an example.

The demo cube I’ve been using in this post has been changed to add a new dimension, called Data Type, which has just one hierarchy with one member on it called Actuals; Data Type is a fairly standard time utility dimension. The Cost % calculation has also been changed so that it’s now a calculated member on the Data Type dimension, although it is still defined on the cube. Here’s its new definition:

CREATE MEMBER CURRENTCUBE.[Data Type].[Data Type].[Cost %] AS

DIVIDE(

([Measures].[Cost Amount],[Data Type].[Data Type].&[Actuals]),

([Measures].[Sales Amount],[Data Type].[Data Type].&[Actuals])),

FORMAT_STRING='0.0%';

Now if I run the following query:

WITH

 

MEMBER [Product].[Product].[Simple Set] AS

AGGREGATE({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]})

 

MEMBER [Product].[Product].[Nextmember Function Used] AS

AGGREGATE({[Product].[Product].&[Apples],

[Product].[Product].&[Apples].NEXTMEMBER})

 

MEMBER [Product].[Product].[Descendants Function Used] AS

AGGREGATE(DESCENDANTS({[Product].[Product].&[Apples],

[Product].[Product].&[Oranges]}))

 

MEMBER [Product].[Product].[Descendants Function Used Twice] AS

AGGREGATE({

DESCENDANTS([Product].[Product].&[Apples]),

DESCENDANTS([Product].[Product].&[Oranges])

})

 

MEMBER [Product].[Product].[Descendants Function Used Twice With Union] AS

AGGREGATE(

UNION(

DESCENDANTS([Product].[Product].&[Apples]),

DESCENDANTS([Product].[Product].&[Oranges])

))

 

SELECT

{[Measures].[Sales Amount]}

*

[Data Type].[Data Type].ALLMEMBERS

ON COLUMNS,

{[Product].[Product].&[Apples],

[Product].[Product].&[Oranges],

[Product].[Product].[Simple Set],

[Product].[Product].[Nextmember Function Used],

[Product].[Product].[Descendants Function Used],

[Product].[Product].[Descendants Function Used Twice],

[Product].[Product].[Descendants Function Used Twice With Union]}

ON ROWS

FROM [Sales With Data Type]

I get these results:

image

Note that for some of the calculations, the Aggregate() function results in a lower solve order in the way we’ve already seen, but not for all of them. Using the NextMember() function, or having two Descendants() functions without wrapping them in a Union() function, seems to stop SSAS assigning the calculation a lower solve order. Ugh. Luckily, though, I have only been able to replicate this with calculated members from two non-measures dimensions; if Cost % is a calculated measure Aggregate() always gives the lower solve order. Apparently this is something that SSAS does on purpose to try to recognise ‘visual total’-like calculated members and make them work the way you want automatically. This is definitely something to beware of if you are using time utility dimensions and calculations on other dimensions though, as it may result in incorrect values being displayed or performance problems if you’re not careful.

[Thanks to Gabi Münster for showing me how Aggregate() works with different sets and Marius Dumitru for confirming that this is intended behaviour]

Written by Chris Webb

October 14, 2014 at 2:30 pm

The DirectSlice Property And Aggregations

with 2 comments

A bit of an obscure one, this, but useful to know nonetheless…

You may or may not be aware of the DirectSlice property, which can be set on a many-to-many relationship in the Dimension Usage pane in Visual Studio:

image

One of the few good descriptions of it comes from this old white paper:

http://technet.microsoft.com/en-us/library/ms345139(v=sql.90).aspx

I quote:

When a measure expression is defined, it may be expensive to compute the result for the all member of the Many-to-Many dimension if the product needs to be computed across an enormous number of rows. And in examples like this, the value for the all member is the simple sum of the fact table. There is one more property on a many-to-many dimension that accommodates this, the Direct Slice. This is a tuple on the Many-to-Many dimension, where the server is instructed to not compute the measure expression except where the cube author guarantees that the aggregate value of the facts matches the result of this expression. For example, in a scenario using measure expression for currency conversion where the sales facts are stored in US Dollars, Currency.USD would be the DirectSlice in the Currency dimension.

Basically, what it does is let you tell SSAS to not evaluate the m2m relationship if a specific selection has been made and if doing so does not change the output of your query. Doing this could improve query performance if evaluating the m2m relationship is expensive – which it often is.

Here’s an example. The screenshot above shows a simplified example based on the Adventure Works cube containing two measure groups, Currency Rate Facts and Internet Sales Facts, and three dimensions, Source Currency, Destination Currency and Date. Destination Currency has a m2m relationship with the Internet Sales Facts measure group that is resolved through the Currency Rate Facts measure group; it also has IsAggregatable=False set and the US Dollar member is the default member.

Now, if you run a query to get data for a measure by year, like this:

select

{[Measures].[Sales Amount]} on 0,

[Date].[Calendar].[Calendar Year].members on 1

from 

[DirectSliceTest]

…and look in Profiler, you’ll see that both the measure groups are being queried by the Storage Engine – as you would expect.

image

What’s more, if you build an aggregation on the Internet Sales Facts measure group at the correct granularity, you’ll see this aggregation being used by the query:

image

Incidentally, there’s a widely-repeated myth that you can’t use aggregations with m2m relationships or measure expressions – this is not true, although you do have to understand the granularity of data that the Storage Engine is querying and building your aggregation in the right place.

Now it’s time to set the DirectSlice property. In this case I set it to a tuple containing US Dollars, which is as I said earlier the default member on the Destination Currency dimension:

([Destination Currency].[Destination Currency].&[US Dollar])

Remember that by setting this property, I’m telling SSAS that it doesn’t need to evaluate the m2m relationship when this member is selected on Destination Currency because the results of any queries will be identical.

If you rerun the query then there are two things to notice in Profiler:

image

First of all, the Currency Rates Facts measure group is no longer being queried, and so the m2m relationship is not being evaluated. This is good. Second, the aggregation seen above is no longer being used. In fact based on my research there’s no way to get an aggregation to be used when the DirectSlice property is set and the m2m relationship is being bypassed (and I’ve tried several variations, even building aggregations that include attributes from the Destination Currency dimension). This is not good. However as soon as you select something other than what is specified in the DirectSlice property, and the m2m relationship is being evaluated, aggregations can be used again.

In summary, then, using the DirectSlice property is a good idea if you have an expensive m2m relationship and you don’t expect your queries to use aggregations. However if you do need to use aggregations to improve query performance you probably should not set the DirectSlice property.

[Thanks to Akshai Mirchandani for helping with this]

Written by Chris Webb

October 6, 2014 at 4:34 pm

Posted in Analysis Services

Counting Customers Who Have Bought All Selected Products

leave a comment »

Market basket analysis is a fairly well understood problem for SSAS – it’s best handled using many-to-many relationships, as shown in the excellent The Many-to-Many Revolution  white paper. However the other week I was asked a question that was an interesting variation on this problem, and which shows up one of the limitations of the many-to-many approach.

Let me show you an example. The following PivotTable on the Adventure Works cube shows the number of distinct customers who bought products in the product category Bikes:

image

The Customer Count measure is a distinct count measure, the Country hierarchy from the Customer dimension is shown on rows, and the slicer here shows the Category hierarchy from the Product dimension. If you select Bikes and Clothing in the slicer, like so:

image

… the measure values are larger because what you’re seeing now is the number of distinct customers who bought products in either the category Bikes or the category Clothing.

But what if you want to see the number of customers who bought Bikes AND Clothing? Well, setting up a many-to-many relationship will give you this (and a lot more useful stuff too) but there are some limitations with this approach:

  • Adding the m2m relationship and the extra measure groups it needs increases processing time
  • You have to select your two product categories from two different (but identical) hierarchies
  • You can only select two product categories – if you want to add a third to your AND filter you need to add a third m2m relationship, and so on. It’s not very flexible.

It is possible to get around these issues with a pure MDX approach, and apply an AND filter that works with however many product categories that the user selects.

The key to understanding how this works is to see how you would write an MDX calculated member that does an AND filter on two product categories:

WITH

MEMBER MEASURES.[Bikes And Clothing] AS

COUNT(

EXISTING 

NONEMPTY(

NONEMPTY(

[Customer].[Customer].[Customer].MEMBERS,

([Measures].[Internet Sales Amount],[Product].[Category].&[3])),

([Measures].[Internet Sales Amount],[Product].[Category].&[1]))

)

SELECT

{MEASURES.[Bikes And Clothing]}

ON 0,

NON EMPTY

[Customer].[Country].[Country].MEMBERS

ON 1

FROM

[Adventure Works]

image

Here I’ve used two nested NonEmpty() functions to filter the set of all customers, first to get the ones that bought Clothing, then to filter these customers again to get the ones that bought Bikes; there’s also an EXISTING there to get only the ones in the current country.

Once again, to add more product categories to the AND filter you need to add more nested NonEmpty() functions… which means you need a way to dynamically generate the code, which of course you can do using StrToSet(). Now normally I avoid using StrToSet() inside MDX calculations because it can cause serious performance problems but in this case it’s the best choice.

The following named set and calculated measure show how to solve the problem for Adventure Works:

CREATE DYNAMIC SET [Selected Product Categories] as  

EXISTING [Product].[Category].[Category].MEMBERS;

 

CREATE MEMBER CURRENTCUBE.MEASURES.[Distinct Customers with all selected Product Categories] AS

IIF(

[Measures].[Internet Sales Amount]=0,

NULL,

COUNT(

EXISTING

STRTOSET(

GENERATE(

[Selected Product Categories],

"NONEMPTY(", "")

+ "[Customer].[Customer].[Customer].MEMBERS" +

GENERATE(

[Selected Product Categories],

", ([Measures].[Internet Sales Amount], " +

[Selected Product Categories].CURRENT.UNIQUENAME

+ "))", "")

)

)

);

The dynamic named set is there to help work out what is selected in the slicer in my Excel worksheet, as shown in Mosha’s old post here. The calculated measure is where all the interesting stuff happens: it uses the ability of the Generate() function to iterate over a set (in this case the dynamic named set), evaluate a string expression and concatenate these strings. The output of this is a series of nested NonEmpty()s, which then goes to StrToSet() to be evaluated, and then the contents of that set are counted.

image

This technique does not replace using a m2m relationship, because it will not allow you to show a query with product category on rows and columns and the number of customers who bought each combination shown. However, if all you want to do is show the number of customers who have bought one specific combination, this has a lot of advantages.

For all of your MDX fans out there, here’s another approach I came up with which doesn’t perform quite as well but is so much fun I had to share it:

CREATE MEMBER CURRENTCUBE.MEASURES.[Distinct Customers with all selected Product Categories V2] AS

IIF(

[Measures].[Internet Sales Amount]=0,

NULL,

COUNT(

UNION(

INTERSECT(EXISTING [Customer].[Customer].[Customer].MEMBERS AS MYCUSTOMERS, {}),

GENERATE(

[Selected Product Categories],

IIF(

[Selected Product Categories].CURRENT.ITEM(0) IS TAIL([Selected Product Categories]).ITEM(0),

NONEMPTY(MYCUSTOMERS, [Measures].[Internet Sales Amount]),

INTERSECT(NONEMPTY(MYCUSTOMERS, [Measures].[Internet Sales Amount]) AS MYCUSTOMERS, {})

)

)

)

)

);

I’m using Generate() here again, but this time I’m using it with inline named sets (which are even worse for performance than StrToSet) in a way similar to this classic post. The theory is the same though: it’s a way of calling NonEmpty() multiple times over a set with an arbitrary number of items in.

Here’s what the calculation does:

  • Inside the first Intersect() I’m declaring an inline named set called MYCUSTOMERS, then doing the Intersect() with an empty set to return… and empty set. But I’ve managed to declare my inline named set, which is the point here.
  • I’m then iterating over the same dynamic named set shown in my earlier example using Generate(), and:
    • For all but the last product category in that set, I’m doing the NonEmpty() on the contents of MYCUSTOMERS on the current product category and then overwriting the contents of MYCUSTOMERS with the output. However, for these iterations of Generate() I’m returning an empty set.
    • For the last product category in the set I’m actually returning the output of NonEmpty() over MYCUSTOMERS for the current product category. The reason I’m only returning something other than an empty set on the last iteration of Generate() is that Generate() returns the union of all the sets returned by each iteration, and that would give me the wrong results! I only want to output the set from the final iteration.

Written by Chris Webb

September 12, 2014 at 9:30 am

Posted in Analysis Services, MDX

Calculating The Value Of Overdue Invoices Using Many-To-Many Relationships in SSAS Multidimensional

with 15 comments

Recently I had to solve the same, rather complex, problem for three different customers within the space of two weeks, a variation of the ‘events-in-progress’ problem I’ve blogged about a lot in the past. It’s this one: how can you calculate the value of your overdue invoices on any given date? It’s important to stress that we are not talking about the invoices that are overdue today – we want to be able to pick any date in the past and find out the value of invoices that were overdue at that point in time.

Let’s imagine you have a fact table containing payments against invoices: each row has an invoice number, the date of the payment, the payment amount, and the date that the invoice is due. A positive value in the Amount column indicates that this is the opening amount of the invoice; negative values in the Amount column are payments against the invoice. Multiple payments can be made against an invoice before it is fully closed. Here’s some example data:

image

Looking at the rows highlighted for invoice 5, you can see in the upper box that there is a positive amount of £35 shown on January 1st 2014 – this is the date that the invoice was opened, and £35 is the full value of the invoice. The invoice is due on January 10th 2014. In the lower box you can see there were four separate payments of £5, £5, £10 and £15 before the invoice was fully paid off on January 12th 2014.

Now, if you were to do a total-to-date in MDX (similar to what I describe here) it would be pretty easy to calculate the outstanding amount on all invoices on any given date. However the requirement here is not only to do that, but to break the value down so that you can see what the total value for overdue invoices and non-overdue (which I’m going to call backlog in this post) invoices is on any given date. This is tricky to do because we have to take the Due Date of each invoice into account as well as the Payment Date. The desired output for invoice 5 is this:

image

As you can see in this PivotTable, for invoice 5 £25 was still outstanding on January 9th 2014, but this is shown as backlog because this is before the due date of January 10th. On January 11th, one day after the due date, the remaining outstanding amount of £15 is shown as overdue. Of course, we also want to be able to calculate the correct values for all invoices:

image

One way of solving this problem would be to calculate the overdue and backlog values for each date that each invoice is open in your ETL, and store these values in a snapshot fact table. This works, and will give you the best possible query performance, but it has two major drawbacks: first, it makes your ETL much slower, and second it makes your fact table much larger. This post describes how you can calculate the overdue and non-overdue values on any given date using many-to-many relationships instead, without having to blow out the fact table.

To go along with the fact table (called FactInvoice) shown above, I have a date dimension table called DimDate (I have deliberately reduced the number of rows here to the dates I have data for, for reasons that will become clear soon):

image

I have a dimension table containing all of the days that an invoice can be overdue or not overdue for (which has one less than double the number of rows as the date dimension table), plus a second column classifying each row as ‘Backlog’ or ‘Overdue’:

image

… and an invoice dimension table that just contains the distinct invoice numbers called DimInvoice.

Now, let me explain how to build the SSAS cube.

Step 1

  • Build dimensions from all of the tables shown above, so you have dimensions called Date, Overdue Days and Invoice.
  • Build a cube with one measure group, based on FactInvoice, and create one measure with AggregateFunction Sum based on the Amount column in that table.
  • Add the Invoice dimension to the cube with a regular relationship. Add the Date dimension to the cube twice, as Payment Date and Due Date, with regular relationships on the PaymentDate and DueDate columns. The Dimension Usage tab should look like this:

image

Step 2

  • Create the following view in SQL Server:
  • CREATE VIEW [dbo].[FactDateToPaymentDate]
    AS
    SELECT        a.DateKey, b.DateKey AS PaymentDateKey
    FROM            dbo.DimDate AS a INNER JOIN
                             dbo.DimDate AS b ON a.DateKey >= b.DateKey

    This view returns all of the combinations of a given date and all dates up to and including the date.

  • Add this view to the DSV and create a new measure group from it; you’ll need to create a measure here, but it can just be a Count measure.
  • Add the Date dimension to the cube once again, this time leaving the name as Date (you will now have three role-playing copies of the Date dimension in the cube).
  • Set up regular relationships between the Date and Payment Date dimensions and the new measure group, then a many-to-many relationship between Date and the Invoice measure group. This means that when you query the cube by the Date dimension, the many-to-many relationship will mean you see the sum of all Amounts whose payment date is up to and including the date selected. The Dimension Usage tab will look like this:

image

Step 4

  • Create another view in SQL Server with this definition:
  • CREATE VIEW [dbo].[FactDueDateToAsOfDate]
    AS
    SELECT        a.DateKey AS DueDateKey, b.DateKey AS AsOfDateKey, DATEDIFF(dd, a.FullDate, b.FullDate) AS OverDueDays
    FROM            dbo.DimDate AS a CROSS JOIN
                             dbo.DimDate AS b

    Yes, I am cross joining the DimDate table with itself and yes, this could return a lot of rows. However you should find that the view is very fast to execute.

  • Add this view to the DSV and build another measure group from it, again with a single Count measure on it.
  • Set up a regular relationship between this new measure group and the Due Date dimension.
  • Add the Overdue Days dimension to the cube, give it a regular relationship with the new measure group and a many-to-many relationship with the Invoice measure group through the new measure group.
  • Add the Date dimension to the cube yet again, creating a new role-playing dimension called As Of Date, give it a regular relationship with the new measure group and a many-to-many relationship with the Invoice measure group through the new measure group. The Dimension Usage tab will now look like this:

image

The purpose of this measure group is this: if you select a date on the As Of Date dimension, you will be able to select ‘Overdue’ on the Overdue Days dimension and this will give you all of the dates on Due Date that were overdue on that date.

Step 5

  • You only want to have to select one date in your PivotTable, so create the following MDX calculated measures that take your selection on the Date dimension and applies it to the As Of Date dimension too:

CREATE MEMBER CURRENTCUBE.MEASURES.[Open Amount Hidden] AS
([Measures].[Amount],
LINKMEMBER([Date].[Date].CURRENTMEMBER, [As Of Date].[Date])
), VISIBLE=FALSE;

CREATE MEMBER CURRENTCUBE.MEASURES.[Open Amount] AS
IIF(MEASURES.[Open Amount Hidden]=0, NULL, MEASURES.[Open Amount Hidden]);

  • Hide all measures apart from the calculated measures you’ve just created
  • Hide the Payment Date, Due Date and As Of Date dimensions

Conclusion

This is a very complex pattern, I know, and this is after I’ve simplified it a lot (if you need currency conversion as well then things get even worse) but I also know it’s extremely useful from a business point of view. Query performance is also reasonably good, at least in the places where I have implemented this.

You can download my sample SQL Server database and VS 2012 project here.

Written by Chris Webb

July 28, 2014 at 10:00 am

Optimising MDX Calculations With The Unorder() Function

with 7 comments

The Unorder() function is probably one of the least used functions in the whole of MDX. It exists only as a query performance hint and, since I had never up to now found a scenario where it did improve the performance of a calculation I had pretty much forgotten about it (as Books Online says, the optimisation it performs is applied automatically in many cases). However I was playing around with some calculations last week and found out that it does have its uses…

What does the Unorder() function do? It’s a function that takes a set and returns a set, and what it does is remove any implicit ordering from that set. By default all sets in MDX are ordered, but for some types of operation that ordering is unimportant and ignoring it can result in faster query performance.

Take, for example, the following query on the Adventure Works cube which shows the number of customers who have bought something up to the current date:

WITH

MEMBER MEASURES.CUSTOMERSTODATE AS

COUNT(

    NONEMPTY(

        [Customer].[Customer].[Customer].MEMBERS

    , {[Measures].[Internet Sales Amount]}

      *

      {NULL : [Date].[Calendar].CURRENTMEMBER})

)

 

SELECT

{MEASURES.CUSTOMERSTODATE}

ON COLUMNS,

[Date].[Calendar].[Date].MEMBERS

ON ROWS

FROM

[Adventure Works]

 

On my laptop it executes in 35 seconds on a cold cache. We can optimise the calculation here simply by wrapping the set of all members on the Customer level of the Customer hierarchy with the Unorder() function, so:

UNORDER([Customer].[Customer].[Customer].MEMBERS)

The following query now executes in 27 seconds on a cold cache:

WITH

MEMBER MEASURES.CUSTOMERSTODATE AS

COUNT(

    NONEMPTY(

        UNORDER([Customer].[Customer].[Customer].MEMBERS)

    , {[Measures].[Internet Sales Amount]}

      *

      {NULL : [Date].[Calendar].CURRENTMEMBER})

)

 

SELECT

{MEASURES.CUSTOMERSTODATE}

ON COLUMNS,

[Date].[Calendar].[Date].MEMBERS

ON ROWS

FROM

[Adventure Works]

 

As far as I can tell, Unorder() only makes a difference on calculations when used in combination with NonEmpty(), and when it is used over a large set (here the set of customers has around 18000 members). If you have calculations like this I would recommend testing to see if Unorder() makes a difference – if it does, please leave a comment and let me know what you find!

Written by Chris Webb

July 7, 2014 at 9:30 am

Using Slicer Selections In The CubeSet Function

with 8 comments

I had an interesting challenge from a customer yesterday – one of those problems that I’d known about for a long time but never got round to working out the solution for…

Consider the following PivotTable, based on a PowerPivot model using Adventure Works data, in Excel 2010:

image

It shows the top 10 products by the measure Sum of Sales. There are two slicers, and the top 10 shown in the PivotTable reflects the selections made in the slicers. All of this works fine. But what if you want to use Excel cube functions to do the same thing? You can write the MDX for the top 10 products quite easily and use it in the CubeSet() function in your worksheet, but how can you get your MDX set expression to respect the selection made in the slicers?

The solution to this problem is very similar to the trick I showed here – finding the selected items in a slicer is not easy! Here are the steps I followed to do it:

  • Add the slicers for EnglishOccupation and CalendarYear to a new worksheet
  • Go to Slicer Settings and uncheck the box for “Visually indicate items with no data”
  • Add two new PivotTables to the worksheet. Connect one to the EnglishOccupation slicer and put EnglishOccupation on rows; connect the other to the CalendarYear slicer and put CalendarYear on rows.
  • Use the OLAPPivotTableExtensions add-in (which you can download here) to add new MDX calculated measures to each PivotTable. For the EnglishOccupation PivotTable call the measure SelectedOccupations and use the following MDX:
    SetToStr(Except(Axis(0), {[Customer].[EnglishOccupation].[All]}))
    This expression does the following: it uses the Axis() function to find the set of members selected on what Excel thinks of as the rows axis in the PivotTable (actually the MDX columns axis), then uses Except() to remove the All Member from the hierarchy (which Excel uses for the Grand Totals) and then uses SetToStr() to take that set and return the string representation of it. Do the same thing for the PivotTable showing CalendarYear too, calling the calculated measure SelectedYears; the MDX in this case is:
    SetToStr(Except(Axis(0), {[Date].[CalendarYear].[All]}))
    This is what the EnglishOccupation PivotTable should look like:
    image
  • Next, to make things easy, use Excel formulas to get the values from the top cell inside each PivotTable into cells elsewhere in the worksheet, and give these cells the names SelectedOccupations and SelectedYears.
    image
  • Then enter a CubeSet() function into a new cell using the following formula:
    =CUBESET(
    "PowerPivot Data",
    "Topcount(
    [Product].[EnglishProductName].[EnglishProductName].members,
    10,
    Sum(" & SelectedOccupations & " * " & SelectedYears & ",[Measures].[Sum of Sales])
    )",
    "Top 10 Set")
    What this does is use the TopCount() function to find the top 10 Products, and in the third parameter of this function which is the numeric expression to find the top 10 by, it crossjoins the two sets of selected occupations and selected years and then sums the output of the crossjoin by the measure [Sum of Sales].
  • Last of all, build your report using the Excel cube functions as normal, using the CubeRankedMember() function to get each item from the top 10 set created in the previous step.

image

You can download my sample workbook here.

The bad news about this technique is that it doesn’t work in Excel 2013 and Power Pivot. It’s no longer possible to create MDX calculated measures on Power Pivot models in Excel 2013, alas. It will work if you’re using any version of Excel from 2007 on against Analysis Services and, as I show here, Excel 2010 and PowerPivot. If you are using Power Pivot and Excel 2013 it might be possible to create a DAX measure to do the same as the MDX I’ve used here (I’m wondering if the technique Jason describes here will work). It would certainly be possible to use CubeRankedMember() to find each item selected in the slicer, as Erik Svensen shows here, and then use Excel formulas to find the MDX unique name for each selected member and concatenate these unique names to create the set expression that my calculated measures return, but that’s a topic for another post. This really should be a lot easier than it is…

Written by Chris Webb

June 20, 2014 at 10:59 am

SSAS Multidimensional Formula Engine Caching and Locale-Dependent Properties

with 4 comments

One subject I have blogged about many times here is how the use of certain MDX functions and features prevents the Formula Engine from caching the result of a calculation for longer than the lifetime of a query (see here and here for just two examples). Reading the new SSAS 2012/2014 Performance Guide, I spotted the following:

The use of MDX functions that are locale-dependent (such as CAPTION or .Properties) prevents the use of the global cache, because different sessions may be connected with different locales and cached results for one locale may not be correct for another locale.

It’s a bit vague but here’s an example of it happening. Take the following calculated measure for the Adventure Works cube:

CREATE MEMBER CURRENTCUBE.MEASURES.TEST AS

IIF(

[Date].[Date].CURRENTMEMBER.PROPERTIES("MEMBER_NAME")="July 1, 2001",

[Measures].[Internet Sales Amount],

[Measures].[Internet Tax Amount]);

If you run the following query, first on a cold cache and then on a warm cache:

select

{measures.test} 

on 0,

[Date].[Date].[Date].members

on 1

from

[Adventure Works]

In Profiler you will see that the second time it is run, the values for the calculated measure are returning from the Formula Engine cache:

image

Here, the Get Data From Cache event is showing that on the second execution of the query the Formula Engine cache is being used.

However, if you change the calculation so it uses the MEMBER_CAPTION property instead of the MEMBER_NAME property, like so:

CREATE MEMBER CURRENTCUBE.MEASURES.TEST AS

IIF(

[Date].[Date].CURRENTMEMBER.PROPERTIES("MEMBER_CAPTION")="July 1, 2001",

[Measures].[Internet Sales Amount],

[Measures].[Internet Tax Amount]);

Then run the query again, on a cold cache then a warm cache, you will see the following:

image

Here the Storage Engine cache is being used – and since the query only contains a single calculated measure, this means that the Formula Engine cache is not being used, and may result in noticeably worse overall query performance if your calculation is expensive.

The reason there is a difference between MEMBER_NAME and MEMBER_CAPTION is that the former property will return the same value for all users, but the latter may return different values for users in different locales. Therefore it is not safe to cache the results of calculations that use the MEMBER_CAPTION property because these calculations could return different values for different users for the same query.

Written by Chris Webb

June 16, 2014 at 9:30 am

Follow

Get every new post delivered to your Inbox.

Join 3,309 other followers