My issue is when I use non empty on the row, all row fields disappear from the data set in Visual Studios 2010. The same happens if I place non empty on columns -- all the column fields disappear from the data set. If the non empty is removed from the statement, all rows and column return to the data set.
This query runs as expected in SQL management studios, but not in Visual Studios. Could someone shed some light on this please?
I've posted just the select query portion below:
SELECT
{
[Quantity - Starting Date Range]
,Receipts
,[Cost of Sales]
,[Samples]
,[Donations]
,[Movement Journals]
,[Subtotal Qty Rollforward]
,[Quantity - Ending Date Range]
,[Qty Rollforward vs AX Difference]
} ON COLUMNS
,NON EMPTY
[Released products].[Product number].[Product number].MEMBERS*
[Released products].[Product name].[Product name].MEMBERS*
[Sites].[Sites - name].[Sites - name].MEMBERS*
[Warehouses].[Warehouses - name].[Warehouses - name].MEMBERS*
[Inventory batches].[Expiration date].[Expiration date].MEMBERS ON 1
FROM [aCube];
NON EMPTY removes all empty(null) values from the selected axes. If you have empty/null values, you won't be seeing them upon using this keyword.
Related
I am facing very strange issue with MDX (SSAS 2014), on which simplest calculated member is taking forever to execute. Could someone please help me to understand why i am facing this issue. If i not use calculated member everything works fast and result comes in seconds. When i remove Producer attribute, query performances well.
Below is the complete query.
WITH
MEMBER Measures.AsOfDate AS ("[Policy Effective Date].[Year-Month].[Date].&[2018-01-04T00:00:00]")
MEMBER Measures.YTDPremium AS AGGREGATE (YTD(STRTOMEMBER(Measures.AsOfDate)), [Measures].[Written Premium])
SELECT NON EMPTY
{
Measures.YTDPremium
} ON COLUMNS, NON EMPTY
{
(
[Program].[Program Name].[Program Name]
,[Insuring Company].[Insuring Company Name].[Insuring Company Name]
,[Line Of Business].[Line Of Business].[Line Of Business]
,[Producer].[Producer Name].[Producer Name]
)
} ON ROWS
FROM [Premium]
Try understand what the following part does in your query
NON EMPTY { ( [Program].[Program Name].[Program Name]
,[Insuring Company].[Insuring Company Name].[Insuring Company Name]
,[Line Of Business].[Line Of Business].[Line Of Business]
,[Producer].[Producer Name].[Producer Name]
) } ON ROWS
In the above MDX you are telling the server to take a cross product of all values of "Programs", "Line Of Business" and "Producer Name". So lets say you have 4 values of programs , 3 values of line of business and 2 values of producer name. The total combinations are 4*3*2=24
Now the "Non empty" removes any combinations that are not present in your dataset. This is done by removing all rows that have "Null" value in column value.
Your measure is returning value irrespective if that combination exists or not. You can modify your Calculatedmeasure to return value only in the case if the combination is valid. This can be achived by checking an actual measure for that combination
Edit: based the below example is based on the comment
In the below example i am trying to get the internet sales amount categories and components
select
{ [Measures].[Internet Sales Amount] }
on columns,
(
[Product].[Category].[Category],
[Customer].[Country].[Country]
)
on rows
from [Adventure Works]
Result
Now add "Non empty" to the query and observe the results.
Results
Now lets add calculted measure that returns "hello". Notice how the non empty clause is ineffective.
Now modify the code make the calculated measure check other measures for null
with member measures.t as
case when [Measures].[Internet Sales Amount] = null then null else "hello" end
select
{ [Measures].[Internet Sales Amount] ,measures.t }
on columns,
non empty
(
[Product].[Category].[Category],
[Customer].[Country].[Country]
)
on rows
from [Adventure Works]
Result
The bottom line: Because of cross product your result is so huge that SSAS is having hard time handling it.
Working on a task that required to bring all a dimension’s data to the Pivot Table(Excel Sheet) even if they are not related to the fact.
first I was able to do it by using the option “PivotTable Options” -> “Display” -> “Show items with no data on rows” from Excel. The problem here is that using this option is going to affect the other dimensions and the requirement is to only this work for the Student dimension only and the user doesn’t like to keep changing the option back every time. then I found this solution using SCOPE, below, but just like above I could not find an away to just have the scope ignore the logic if any other dimension added so the data does not get duplicated.
SCOPE ([Program].[Program Hierarchy].MEMBERS, [Measures].[Number of Students]);
THIS = IIF(ISEMPTY([Measures].[Number of Students]), 0, ([Measures].[Number of Students]));
END SCOPE;
So is there something that I can add the SCOPE above to just work in the program dimension and get ignore/ skip and work as normal if any other dimension added to the pivot table?
Any suggestion will be appreciated.
Lets suppose that for Dimension1 ,attribute1, and the attribute "Value 1", you had nothing in Fact, so this is removed from the result, now you forced Excel to display it by selecting the option. When you add another dimension's attribute lets say Dimension2.Attribute1, since "Value 1" had nothing in Fact, therefore Cube will not understand which value of of Dimension2.Attribute1 is to be displayed in front of Dimension1.Attribute1, therefore it will display all its values. So if we have 3 values in Dimension2, attribute 1 then "Value 1" will be repeated three times. Now with Excel you cannot solve the issue, however it might just be possible to write an MDX query that works.
Edit: Query Added.
The below sample query is based on AdventureWorks, the first sample shows that the result has some nulls, if i un-comment the "non-empty" all null values will vanish, go ahead try it.
select [Measures].[Internet Sales Amount] on columns,
--non empty
[Product].[Subcategory].[Subcategory]
on rows
from
[Adventure Works]
Result without non empty
Now lets add another dimension to the query. Notice that the null value for the first row(Bib-shorts) is now repeated for all values of the second dimension, Since cube has no way to determine which value to display.
select [Measures].[Internet Sales Amount] on columns,
--non empty
([Product].[Subcategory].[Subcategory],[Date].[Calendar Quarter of Year].[Calendar Quarter of Year])
on rows
from
[Adventure Works]
Result
Now the above result shows the issue you are facing. What we now need to do is whenever there is a null value we dont need the individual members of the second dimension, rather a place holder to satisfy the tuple, will work.
In the query below I have two tuples
1) for the not null data-points. Here we display the actual member of the second dimension.
2) for the null data-points, here we use ".defaultmember" which basically means that the second dimension will behave as it was never selected. Have a close look at the second dimension it says "All Period"
select [Measures].[Internet Sales Amount] on columns,
--non empty
{filter(([Product].[Subcategory].[Subcategory],[Date].[Calendar Quarter of Year].[Calendar Quarter of Year]),[Measures].[Internet Sales Amount]>0),
filter(([Product].[Subcategory].[Subcategory],[Date].[Calendar Quarter of Year].defaultmember),[Measures].[Internet Sales Amount]=null)
}
on rows
from
[Adventure Works]
Result:
WITH
MEMBER CostDifference AS
Sum([Measures].[ExtendedCost]) - [Measures].[ExtendedCost]
SELECT
NON EMPTY
{
[Measures].[ExtendedCost]
,CostDifference
} ON COLUMNS
,NON EMPTY
{[Parts].[ItemDesc].MEMBERS, [Dim Date].[DateUK].MEMBERS} ON ROWS
FROM [Cube]
I'm trying to subtract a measures across different dimensions i.e. for 2 date snapshots (31/1/2010 and 28/2/2010) for all the products in DimParts
if I remove this piece of code from equation then all I get is zeros
[Dim Date].[DateUK].MEMBERS
If I use a cross join to add more than one dim even then the cost difference is zero
e.g. ,NON EMPTY
CrossJoin
(
[Parts].[ItemDesc].MEMBERS
,{[Dim Date].[DateUK]}
) ON ROWS
I'm using SQl Server 2008R2
What am I missing here.
This Sum([Measures].[ExtendedCost]) - [Measures].[ExtendedCost]
Will resolve to this
[Measures].[ExtendedCost] - [Measures].[ExtendedCost]
Which is always 0
If this Sum([Measures].[ExtendedCost]) needs to be across a complete set then you need to include that set:
Sum(
[Dim Date].[DateUK].MEMBERS
,[Measures].[ExtendedCost]
)
I have a requirement displaying data from same dimension in more than 1 column. For eg. I want to show data Year and Month wise. In my dimension structure, Year and Month belongs to same hierarchy. When I run below query I get error. PFB the query.
Select NON EMPTY {[Measures].[Target Actual Value]} ON 0,
NON EMPTY {[Realization Date].[Hierarchy].[Year Name].Members *
[Realization Date].[Hierarchy].[Month Year]} ON 1
From [Cube_BCG_OLAP]
The error I get is Query (2, 12) The Hierarchy hierarchy is used more than once in the Crossjoin function. I am new to MDX queries. Please help in this regard. Thanks in advance.
Select NON EMPTY {[Measures].[Target Actual Value]} ON 0,
NON EMPTY {[Realization Date].[Hierarchy].[Year Name].Members ,
[Realization Date].[Hierarchy].[Month Year]} ON 1
From [Cube_BCG_OLAP]
Instead of CROSSJOIN have a set as above. In a set, you can put members from same hierarchy
I like Sourav's answer - but it will put the results in one column which is slightly different than the question.
In AdvWorks this is in one column:
SELECT
[State-Province].MEMBERS ON COLUMNS
,{
[Date].[Calendar].[Calendar Year].MEMBERS
,[Date].[Calendar].[Month].MEMBERS
} ON ROWS
FROM [Adventure Works];
It is possible to switch to two columns and use a cross join but you need to find out the details of your Date dimensions Attribute hierarchies (as opposed to User hierarchies):
SELECT
[State-Province].MEMBERS ON COLUMNS
,
[Calendar Year].[All Periods].Children
* [Month].MEMBERS ON ROWS
FROM [Adventure Works];
In your cube maybe something like this:
SELECT
NON EMPTY
{[Measures].[Target Actual Value]} ON 0
,NON EMPTY
[Year Name].MEMBERS
*
[Month Year].MEMBERS ON 1
FROM [Cube_BCG_OLAP];
Simple AdwWorks cube has time dimension with years range from 2000-th to 2007-th. Fact data exists for 2001-2006. 2000-th and 2007-th are empty. Ok.
I'm insert new data to source fact table for 2001-th and process the cube. All changes immediately reflect in measure value (it grows). But new data for 2000-th don't reflect in measure value growing. Nevertheless, during cube processing SSAS sees new rows for both cases (the value of rows count grows), and sql profiler catches equal batches of commands for 2001-th and 2000-th. But measure value grows only for 2001-th.
I've cleared mdx-script - it contains now only calculate command. Now there is:
a) select sum(measure value) on fact tables in source database reflect new rows for both cases
b) cube processing sees changes for both cases
c) mdx-script is clear (there are no scopes, witch can set measure value to null )
d) select [sales amount] on cube reflects this changes only for 2001-th, not for 2000-th
Any suggestions?
================================================
One day later.. :)
The problem seems as resolved, but... I don't know: is it tech and/or logic feature or bug..
So, what is happened.. Simple mdx query:
select [sales amount] on 0,
[some date from 2000] on 1
from [cube]
gives null value for [sales amount], but for [some date from 2001] value is not empty. In database fact table both values (results of sum(salesamount) for both dates) are not empty. After research i found that there was one measure ([Temp]) in the [sales amount] measure group for which MeasureExpression was set to [sales amount]*[average rate] ([average rate] - measure from [Fact Currency Rate] measure group). Fact table for [Fact Currency Rate] measure group didn't contain data for 2000-th. I've insert it and - vualya - all work and for 2000-th now.
Main question is: why ssas calculate MeasureExpression formulas for measures, which I'm not request in a query, and why ssas set requested by me value to null, if it does not depend from formula of MeasureExpression property of unrequested measure?
Is it bug or "defense from fool"?. There is another sample of such strange ssas behaviour. I mean usage of UnaryOperatorColumn property. If it is set for any attribute of any dimension, then it influences on every query, even if this attribute not is a part of this query. For example
select [sales amount] on 0
from [cube]
returns different results, which depends by setting of UnaryOperatorColumn for some dimension attribute..
What all it mean?