Tag Archives: MDX

MDX, Pragmatic Works, SSAS

Ten MDX Calculations For Your Cube

8 Comments

I get lots of questions about how to implement various types of calculations in a cube so I figured I’d cover some of the more commonly used calculations I run into on a regular basis. All of these example calculations I’ll be providing were written against the Adventure Works 2008 R2 example cube but these should still work against later versions of the AW cube.

10. YTD, QTD, and MTD Calculations

These kinds of calculations are pretty common and I see these in a lot of cubes. They’re pretty easy to wire up since there are only a couple simple MDX functions necessary to make this work.

CREATE
MEMBER CurrentCube.[Measures].[YTD Internet Sales Amount] AS
Aggregate
(
PeriodsToDate
(
[Date].[Calendar].[Calendar Year]
,[Date].[Calendar].CurrentMember
)
,[Measures].[Internet Sales Amount]
)
,FORMAT_STRING = “Currency”
,VISIBLE = 1
,DISPLAY_FOLDER = ’10 MDX Calculations’
,ASSOCIATED_MEASURE_GROUP = ‘Internet Sales’;

Let’s start with the PeriodsToDate function. The PeriodsToDate function is going to return all the sibling members at the specified level up to the current member. The outer function, the Aggregate function, calculates the value based on the aggregation type specified in the cube for the measure. In our case, the Internet Sales Amount aggregation type is Sum, so the Aggregate function could be exchanged for the Sum function.

If I wanted to calculate the Month to Date for Internet Sales, I would simply exchange the reference for the Calendar Year level of the Calendar hierarchy with the Month level, as seen here:

CREATE
MEMBER CurrentCube.[Measures].[MTD Internet Sales Amount] AS
Aggregate
(
PeriodsToDate
(
[Date].[Calendar].[Month]
,[Date].[Calendar].CurrentMember
)
,[Measures].[Internet Sales Amount]
)
,FORMAT_STRING =
Currency
,VISIBLE = 1
,DISPLAY_FOLDER = ’10 MDX Calculations’
,ASSOCIATED_MEASURE_GROUP = ‘Internet Sales’;

9. Period over Period Growth

This is also a pretty common calculation I find and implement for my clients. It’s also pretty straight forward. In this example, we’re comparing this year’s Internet Sales Amount to the previous year’s Internet Sales Amount.

Create Member CurrentCube.[Measures].[Yearly Growth Internet Sales Amount]

As
([Date].[Calendar Year].CurrentMember,[Measures].[Internet Sales Amount])-
([Date].[Calendar Year].PrevMember,[Measures].[Internet Sales Amount]),

FORMAT_STRING = “Currency”,
VISIBLE = 1 ,  DISPLAY_FOLDER = ’10 MDX Calculations’ ,  ASSOCIATED_MEASURE_GROUP = ‘Internet Sales’  ;

The function here to pay attention to is the PrevMember function. The PrevMember function returns the previous member at a given level based on the supplied member. For example, if the supplied member was 2011, the previous member would be 2010.

Create Member CurrentCube.[Measures].[Monthly Growth Internet Sales Amount]

As
([Date].[Month of Year].CurrentMember,[Measures].[Internet Sales Amount])-
([Date].[Month of Year].PrevMember,[Measures].[Internet Sales Amount]),

FORMAT_STRING = “Currency”,
VISIBLE = 1 , DISPLAY_FOLDER = ’10 MDX Calculations’ , ASSOCIATED_MEASURE_GROUP = ‘Internet Sales’ ;

Similar to our last calculation, if we wish to calculate the growth from month to month, we should reference the Month attribute of our Date dimension.

8. Profit Margin Calculation

Another popular calculation for obvious reasons is the calculation for profit margin. Basically all you need to understand for this calculation is the logic of a Case statement. We use the Case statement to check for a zero value in the denominator.

CREATE
MEMBER CurrentCube.[Measures].[Internet Profit Margin] AS
CASE
WHEN
IsEmpty([Measures].[Internet Sales Amount])
THEN NULL
ELSE
([Measures].[Internet Sales Amount]
–
[Measures].[Internet Total Product Cost])
/
[Measures].[Internet Sales Amount]
END
,FORMAT_STRING = “Percent”
,VISIBLE = 1
,DISPLAY_FOLDER = ’10 MDX Calculations’
,ASSOCIATED_MEASURE_GROUP = ‘Internet Sales’;

7. Percent of Total

More often then not, my clients want to be able to calculate the percent of a total amount for a hierarchy. In this example, I’m once again leveraging the Case statement.

CREATE
MEMBER CurrentCube.[Measures].[Percent of Internet Product Sales] AS
CASE
WHEN
IsEmpty([Measures].[Internet Sales Amount])
THEN NULL
ELSE
(
[Product].[Product Categories]
,[Measures].[Internet Sales Amount]
)
/
(
[Product].[Product Categories].[All]
,[Measures].[Internet Sales Amount]
)
END
,FORMAT_STRING = “Percent”
,VISIBLE = 1
,DISPLAY_FOLDER = ’10 MDX Calculations’
,ASSOCIATED_MEASURE_GROUP = ‘Internet Sales’;

To adapt this calculation to your cube, just replace [Product].[Product Categories] with your dimension and hierarchy.

6. Count Leaf Members of a Hierarchy

I’ve seen this kind of calculation used a lot in combinations with other measures and/or calculation. This calculation counts the Leaves, which are products, of the Product Categories hierarchy.

CREATE
MEMBER CurrentCube.[Measures].[Product Count] AS
Count
(
Descendants
(
[Product].[Product Categories].CurrentMember,
,LEAVES
)
)
,FORMAT_STRING = “#,##0.00;-#,##0.00”
,VISIBLE = 1
,DISPLAY_FOLDER = ’10 MDX Calculations’
,ASSOCIATED_MEASURE_GROUP = ‘Internet Sales’;

There’s not much to this calculation. First we use the Descendants function with the Leaves function as the third argument to get all the Leaf members of the Product Categories hierarchy. Then we use the Count function to count those members.

While these calculations are not extremely complex, MDX can be a bit overwhelming. A very helpful tool that I’ve found useful, even as someone experienced with MDX, is the MDX Calculation Builder included with BI xPress. Adding calculations, like those seen above, is very easy and only take a few click with the wizard.

To open the MDX Calculation Builder, click the icon seen on the Calculations tab of your Cube Designer in BIDS.

10-30-2011 10-22-07 PM

After you click the Calculation Builder icon, the Calculation Builder wizard will open. The first thing you’ll need to do is select a calculation. At these step, we can also choose to create a named set, which the wizard will help us do.

10-30-2011 10-29-37 PM

After selecting the calculation, you’ll be taken through a couple more steps, such as selecting the Measure you wish to perform the calculation with or specifying which attribute to use as the Year attribute for the calculation.

You’ll also have the opportunity to apply conditional formatting to your measure values, which is a nice touch the end users like to see.

10-30-2011 10-46-34 PM

Finally, we’ll give our measure a name, select the measure group we’d like the measure to be a part of, select the formatting, and preview the calculation that was written by the tool.

Then we’ll click Finish. The calculation has been added to our calculation script within our cube.

10-30-2011 10-50-58 PM

BI xPress has made adding calculations to your cube simply and fast and I take advantage of the tool anytime I’m implementing calculations into my client’s cube. Head to PragmaticWorks.com to check out BI xPress and to check out some of the great training videos hosted there covering the in’s and out’s of MDX.

Stay tuned for my next blog post where we will cover the remaining 5 MDX calculations to add to your cube!

Free Training, MDX, SSRS

SQL Saturday #86: BI Edition in Tampa, FL – Reporting on a Cube w/ SSRS 2008

Next month on November 5th is SQL Saturday #86 in Tampa, Florida. But this is no ordinary SQL Saturday. This is SQL Saturday: BI Edition! Which means this SQL Saturday is going to be even more amazing, hard core, and in your face than usual! If you’ve never been to a SQL Saturday event, let me just say that you are missing out. And if you have been to one before, then you know exactly how awesome these events are. Whether you show up for the incredible FREE training from qualified industry pro’s, the networking with fellow SQL Server fiends, or the great lunch and snacks, you’re sure to have a good time.

So before you waste anymore time, head on over to their site and get registered. Check out the extensive list of sessions and speakers and plan your day.

Yours truly will be giving a session covering Reporting on a Cube w/ SSRS 2008 at 10 a.m.  I’ll be covering some of the advantages the cube presents as a data source and how you can leverage the power of the cube with SQL Server Reporting Services 2008. So head on over to SQLSaturday.com and get registered before it’s completely full! We’re going to have a blast!

MDX, SSAS, SSRS

Showing an SSAS Hierarchy as Delimited String

I recently encountered a requirement for a client to display a list of all the individual stores of a ragged hierarchy (the leaf members) on an SSRS report with a bunch of different measures. But they also wanted to be able to quickly see the the ancestor members of each level above each store quickly by using a tool tip to display the different ancestors.

Turns out there are a couple nifty, little MDX functions that can allow us to do this very easily. The Ancestors and Generate functions make this possible. Here an example I put together with the Adventure Works 2008 R2 cube so you can play along.

WITH MEMBER [Measures].[Management Structure] AS
            GENERATE(
            EXCEPT(({Ancestors([Employee].[Employees].CURRENTMEMBER,6),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,5),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,4),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,3),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,2),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,1)}),[Employee].[Employees].[All]),
            [Employee].[Employees].CURRENTMEMBER.Name, "  >  ")

Select {[Measures].[Management Structure],
    [Measures].[Reseller Sales Amount],
    [Measures].[Reseller Order Quantity]} on 0,

NON EMPTY(DESCENDANTS([Employee].[Employees],10,LEAVES)) ON 1

From [Adventure Works]

Where [Employee].[Title].&[Sales Representative]

Pay attention to the calculate member [Measures].[Management Structure]. The way the Generate function works is that it takes one set and applies that set to each Member in another set. In this case, our first set is the ancestor from each level in our hierarchy of the current member, excluding the All member.

So this our first set:

EXCEPT(({Ancestors([Employee].[Employees].CURRENTMEMBER,6),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,5),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,4),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,3),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,2),
                  Ancestors([Employee].[Employees].CURRENTMEMBER,1)}),[Employee].[Employees].[All])

Our second set is the Name property of the CurrentMember of the Employee hierarchy being displayed:
 
[Employee].[Employees].CURRENTMEMBER.Name
 
The Generate function actually also excepts a third argument: A delimiter. We can specify a delimiter. In my example, I use “  >  “ to make the string a bit easier to read. If we execute the query we can see the results of our calculated member.

10-10-2011 9-47-39 PM

Pretty cool, huh? So now that we have each employees management hierarchy in sweet, little string, we can easily display this in a report as a tool tip.

I added the tool tip to the cell in my report’s tablix that contains the employee name. Here’s the expression I used:

=REPLACE(Fields!Management_Structure.Value,">",VBCRLF)

This expression replaces the “>” with a carriage return, line feed. Here’s what our tool tip looks like:

10-10-2011 10-02-24 PM

Using this method, we can easily display the managers for each employee in a tooltip without having to waste any real estate with additional columns.

I hope somebody finds this useful. I’m out!

MDX, SSAS

Daily Average Orders Calculation at Any Level of Date Hierarchy

1 Comment

Today I was helping a fellow BIDN.com user come up with a calculation to figure out the average numbers of orders on day at any level of a date hierarchy. So I figured I’d post the calculation because 1) maybe it will be helpful to you and 2) I’m going to forget this calculation so I’ll at least be able to use this post as a reference.

WITH 
    //Gets a Summation of the Orders in the Fiscal period
    MEMBER [Measures].[Sum] AS 
    
        SUM([Date].[Fiscal].CHILDREN,[Measures].[Internet Order Count])
    //Counts the days in the Fiscal period
    MEMBER [Measures].[Count] AS 
    
        COUNT({DESCENDANTS([Date].[Fiscal].CURRENTMEMBER, [Date].[Fiscal].[Date])})
    //Divides the sum of orders by the count of days to calculate the average daily 
    //orders in that fiscal period
    MEMBER [Measures].[AvgDailyIntOrdCount] AS  
    
        SUM([Date].[Fiscal].CHILDREN,[Measures].[Internet Order Count])/
        COUNT({DESCENDANTS([Date].[Fiscal].CURRENTMEMBER, [Date].[Fiscal].[Date])})
        
SELECT     {[Measures].[Sum],[Measures].[Count],[Measures].[AvgDailyIntOrdCount]} on 0,    
{[Date].[Fiscal].Members} ON ROWS 
FROM
      [Adventure Works]

In the calculation, I use the SUM function to get an aggregations of the Internet Order Count for any level of the Fiscal Date hierarchy. I also used the Count function with the Descendants function to get a count of days beneath a member of any level of the Fiscal Date hierarchy.

It seems to work pretty good, but if you have an alternative way to calculate this, please post it in the comments section!