I have the following MDX statement on the $Monitoring cube:
WITH
MEMBER [Measures].[Unique Users] AS distinctcount(([User].[User].[User L].members
, [Measures].[Open Report Count])),format_string="#,#0"
SELECT
NON EMPTY { {[Measures].[Unique Users],[Measures].[Open Report Count]} } ON COLUMNS,
NON EMPTY { [Time].[Time].[Day L] } ON ROWS
FROM ( SELECT
{ lag(strtomember("[Time].[Time].["+right("0"+str(day(SchemaRefreshTime())),2) +"-"+ right("0"+str(month(SchemaRefreshTime())),2) + "-"+str(year(SchemaRefreshTime()))+ "]"), 6):strtomember("[Time].[Time].["+right("0"+str(day(SchemaRefreshTime())),2) +"-"+ right("0"+str(month(SchemaRefreshTime())),2) + "-"+str(year(SchemaRefreshTime()))+ "]") } ON 0 FROM [$Monitoring])
/*ic3navigation*/
axis 1 NON EMPTY order(nonempty(Descendants([Report].[Report], ,leaves),[Open Report Count]),[Open Report Count],desc)
FILTERBY /*ic3*/ {[Time].[Time].[ALL].&[2015].&[2015-11-27].&[27-11-2015]}
FILTERBY /*ic3*/ {[User].[User].[All Users].&[<user>]}
*) change <user> with the actual user name
*) the ...lag.. formula is used to give the last 7 days based on schema refresh time
***) this MDX query can be run on any $Monitoring cube if you have filled in an existing user
I would expect the distinctcount function to take into account the FILTERBY. So the result should be 1 (there is just one user selected). The strange thing is, that it does not. It shows more than one user, so I assume the FILTERBY on users is not taken into account for the distinctcount.
The same thing happens when I move the FILTER BY to the AXIS or to the ROWS or COLUMNS.
Is this a bug or is this something how MDX/ MDX++ works in icCube?
Please advise.
It's the expected behaviour. Welcome to advanced MDX !
A FilterBy is exactly the same as a sub-select.
Members are not filtered by a subselect or a where clause in a calculated members.
In a calculated member, a tuple that defines a hierarchy will 'overwrite' the one defined in a subquery or where clause.
------ UPDATE ------
If you want to filter the set with the where clause/subselect you've the EXISTING operator.
MEMBER [Measures].[Unique Users] AS count( Existing [User].[User].[User L].members),format_string="#,#0"
if you want only users with data for the cell tuple :
MEMBER [Measures].[Unique Users] AS count( nonempty( Existing [User].[User].[User L].members, [Measures].[Open Report Count])),format_string="#,#0"
If you've a large number of users I'd advise adding a measure [Distinct Users] that is a distinct count on the user id. This way you avoid all complexity that we're facing here.
Related
I raised a ticket on grand totals when filtering ages ago and I used the proposed solution yesterday to get one dimension working:
WITH
SET [FilteredSet] AS
{
FILTER([CDR].[CDR].Members,([CDR].[CDR].CURRENTMEMBER.MEMBERVALUE="70648") OR ([CDR].[CDR].CURRENTMEMBER.MEMBERVALUE="70646"))
}
MEMBER [CDR].[CDR].[Test] AS Aggregate([FilteredSet])
SELECT NON EMPTY {[Measures].[CS01.TenorTotal] } ON COLUMNS, NON EMPTY {[FilteredSet], [CDR].[CDR].[Test]} ON ROWS FROM [TraderCube] WHERE ([Date].[Date].[2022-04-01],[Epoch].[Epoch].[Branch].[master].[17])
(Other ticket with useful info)
MDX - Grand total missing when filtering
And as I expand the test query (to get it all working and I then expand the C# code to add it) I note that this works:
WITH
SET [FilteredSet] AS
{
FILTER(CROSSJOIN(HIERARCHIZE(DRILLDOWNLEVEL([CDR].[CDR].[ALL].[AllMember]),POST),HIERARCHIZE(DRILLDOWNLEVEL([Book].[Book].[ALL].[AllMember]),POST)),(LEFT([CDR].[CDR].CURRENTMEMBER.MEMBERVALUE,2)="66") )
}
SELECT NON EMPTY {[Measures].[CS01.TenorTotal] } ON COLUMNS, NON EMPTY {[FilteredSet]} ON ROWS FROM [TraderCube] WHERE ([Date].[Date].[2022-04-01],[Epoch].[Epoch].[Branch].[master].[17])
But we will often filter on the 2nd (or 3rd, 4th etc) dimension, like so and the total (AllMember) is filtered out from Book as it does match the INSTR comparisons.
How do I get my Aggregate MEMBER to work properly here? I've placed [?] in position as I can get this sort of setup to work on one dimension (E.g. [CDR] as shown in first code block) but here the query returns [CDR] and then the [Book] members that match and I want to 'glue' my Agregate MEMBER [?] of the measure as a grand total.
WITH
SET [FilteredSet] AS
{
FILTER(CROSSJOIN(HIERARCHIZE(DRILLDOWNLEVEL([CDR].[CDR].[ALL].[AllMember]),POST),HIERARCHIZE(DRILLDOWNLEVEL([Book].[Book].[ALL].[AllMember]),POST)),(LEFT([CDR].[CDR].CURRENTMEMBER.MEMBERVALUE,2)="66")
AND (INSTR([Book].[Book].CURRENTMEMBER.MEMBERVALUE,"72")>0 OR INSTR([Book].[Book].CURRENTMEMBER.MEMBERVALUE,"73")>0))
}
MEMBER [?] AS Aggregate([FilteredSet])
SELECT NON EMPTY {[Measures].[CS01.TenorTotal] } ON COLUMNS, NON EMPTY {[FilteredSet], [?]} ON ROWS FROM [TraderCube] WHERE ([Date].[Date].[2022-04-01],[Epoch].[Epoch].[Branch].[master].[17])
Current MDX error back is:
'The 2 objects do not have the same dimensionality: [[CDR], [Book]] and [[Measures]]'
Thanks to more advanced MDX-ers out there! :)
Leigh Tilley
TilleyTech Ltd
P.S. Sorry no screenshots...client browser trackers would detect it and I'd get in trouble!
I have one dimension that I want to put into filter, and created calculated member that should dynamically show number of selected members from the dimension.
The dimension does not have an All member.
So this is my attempt
with member [Measures].[Count1] as count(existing(([MyDimension].[MyDimensionHierarchy].members)))
select [Measures].[Count1] on 0
from [MyCube] -- gives me 1
and this one will give me 2 which is correct:
with member [Measures].[Count1] as count(existing(([MyDimension].[MyDimensionHierarchy].members)))
select [Measures].[Count1] on 0
from [MyCube]
where ({[MyDimension].[MyDimensionHierarchy].[Member1], [MyDimension].[MyDimensionHierarchy].[Member2]})
But, the problem is that when I create calculated member with the formula above, and drag Count1 to the Excel pivot table, and drag MyDimension as filter, and when I do multi-select of the dimension members, I want the count to dynamically change as I change number of members that are selected.
But Count1 always stays equal to 1.
In a meantime I have found an answer:
The query that I wrote in the question actually is not the query that Excel pivot table sends to the cube. Excel pivot table generates query like this:
SELECT FROM (SELECT ({[MyDimension].[MyDimensionHierarchy].[Member1],[MyDimension].[MyDimensionHierarchy].[Member2]}) ON COLUMNS
FROM [MyCube])
WHERE ([Measures].[Count1])
The way this should be done is by using dynamic set that contains filtered members:
create dynamic set [SelectedMembers] as existing( [MyDimension].[MyDimensionHierarchy].members )
And then:
create member Measures.SelectedMembersCount as count([SelectedMembers])
So this set dynamically changes as different members are selected in the filter and SelectedMembersCount is dynamically changed along the way.
I have a question related to creating a (more efficient) custom Distinct Count Measure using MDX.
Background
My cube has several long many to many relationship chains between Facts and Dimensions and it is important for me to be able to track which members in certain Dimensions do and do not relate to other Dimensions. As such, I have created a "Not Related" record in each of my dimension tables and set those records' ID values to -1. Then in my intermediate mapping fact tables I use the -1 ID to connect to these "Not Related" records.
The issue arises when I try to run a normal out-of-the-box distinct count on any field where the -1 members are present. In the case that a -1 member exists, the distinct count measure will return a result of 1 more than the true answer.
To solve this issue I have written the following MDX:
CREATE MEMBER CURRENTCUBE.[Measures].[Provider DCount]
AS
//Oddly enough MDX seems to require that the PID (Provider ID) field be different from both the linking field and the user sliceable field.
SUM( [Providers].[PID Used For MDX].Children ,
//Don't count the 'No Related Record' item.
IIF( NOT([Providers].[PID Used For MDX].CURRENTMEMBER IS [Providers].[PID Used For MDX].&[-1])
//For some reason this seems to be necessary to calculate the Unknown Member correctly.
//The "Regular Provider DCount Measure" below is the out-of-the-box, non-MDX measure built off the same field, and is not shown in the final output.
AND [Measures].[Regular Provider DCount Measure] > 0 , 1 , NULL )
),
VISIBLE = 1 , DISPLAY_FOLDER = 'Distinct Count Measures' ;
The Issue
This MDX works and always shows the correct answer (yeah!), but it is EXTREMELY slow when users start pulling Pivot Tables with more than a few hundred cells that use this measure. For less than 100 cells, the results are nearly instantaneously. For a few thousand cells (which is not uncommon at all), the results could take up to an hour to resolve (uggghhh!).
Can anyone help show me how to write a more efficient MDX formula to accomplish this task? Your help would be GREATLY appreciated!!
Jon Oakdale
jonoakdale#hotmail.com
Jon
You can use predefined scope to nullify all unnecessary (-1) members and than create your measure.
SCOPE ([Providers].[PID Used For MDX].&[-1]
,[Measures].[Regular Provider DCount Measure]);
THIS = NULL;
END SCOPE;
CREATE MEMBER CURRENTCUBE.[Measures].[Provider DCount]
AS
SUM([Providers].[PID Used For MDX].Children
,[Measures].[Regular Provider DCount Measure]),
VISIBLE = 1;
By the way, I used in my tests [Providers].[PID Used For MDX].[All].Children construction since don't know, what is dimension / hierarchy / ALL-level in your case. It seems like [PID Used For MDX] is ALL-level and [Providers] is name of dimension and hierarchy, and HierarchyUniqueName is set to Hide.
I need to create the table of the following structure in MDX (to be used in SSRS report):
For that I have 2 dimensions and one measure:
Option dimension, with option type and option value attributes
Standard dimension, with IsStandard flag
Price measure
In first column I need to show all option type attributes,
in second all value attributes where IsStandard flag is set to [Y],
in third values chosen by user in parameters and
in fourth prices for components selected by user.
Is it possible to do the above in single MDX? Or will I be better off creating 2 distinct queries and creating 2 tables for them?
EDIT: Since my updates won't fit into the comment, I will add some thoughts here.
EXISTS function from answer below does not filter the result set, I don't get standard values but all possible values concatenated. When I issue the following code:
SELECT
[Measures].[Price] ON 0,
NON EMPTY [Option].[Option Type].children
*
[Option].[Option Value].children ON 1
FROM [Cube]
WHERE
(
[Standard].[IsStandard].&[Y],
[Configurations].[Configuration].&[conf1]
)
It returns the default values correctly, but if I use
SELECT
[Measures].[Price] ON 0,
[Option].[Option Type].children
*
EXISTS(
[Option].[Option Value].[Option Value].members
,([Standard].[IsStandard].&[Y],[Configurations].[Configuration].&[conf1])
) ON 1
FROM [Cube]
It does not filter the results.
If you can accept a slightly different order of columns, then this can be done in MDX, using a calculated measure which is actually a string (as you want to see a list of attributes values in column). This avoids having the same attribute twice in the rows:
WITH Member Measures.[Standard Value] AS
Generate(NonEmpty([Option].[Option Type].[Option Type].Members,
{([Standard].[IsStandard].&[Y],
Measures.[Price]
)}
),
[Option].[Option value].CurrentMember.Name,
", "
)
SELECT { Measures.[Standard Value], Measures.[Price] }
ON COLUMNS,
NON EMPTY
[Option].[Option Type].[Option Type].Members
*
{ #chosenValues } // the parameters value should be a comma separated list like "[Option].[Option value].[AMD], [Option].[Option value].[INTEL]"
ON ROWS
FROM [Your Cube]
WHERE [Configurations].[Configuration].&[conf1]
You can adapt the list separator (the last argument of the Generate function) to anything you like.
And in case there is more than one measure group that is related to the dimensions [Option], [Standard], and [Configurations], you should add the name of the measure group to use for determining the relationship as additional last parameter to the Exists, so that you and not the engine determines that. Just use the name of the measure group in either single or double quotes.
Yes it is, dimension will just be ignored. This is assuming you've all in the same schema / cube.
Note, depending on the OLAP Server you're using it's possible you've to change a flag that sends an error if you're using a dimensions that is not defined at Measure Group level.
Say I have a measure, foo, in a cube, and I have a reporting requirement that users want to see the following measures in a report:
total foo
total foo excluding instances where foo > 10
total foo excluding instances where foo > 30
What is the best way to handle this?
In the past, I have added Named Calculations which return NULL if foo > 10 or just foo otherwise.
I feel like there has to be a way to accomplish this in MDX (something like Filter([Measures].[foo], [Measures].[foo] > 10)), but I can't for the life of me figure anything out.
Any ideas?
The trick is that you need to apply the filter on your set, not on your measure.
For example, using the usual Microsoft 'warehouse and sales' demo cube, the following MDX will display the sales for all the stores where sales were greater than $2000.
SELECT Filter([Store].[Stores].[Store].members, [Unit Sales] > 2000) ON COLUMNS,
[Unit Sales] ON ROWS
FROM [Warehouse and Sales]
I met similar problem when use saiku (backend with Mondrain), as I haven't found any clear solution of "add filter on measure", I added it here, and that may be useful for other guy.
In Saiku3.8, you could add filter on UI: "column"->"filter"->"custom", then you may see a Filter MDX Expression.
Let's suppose we want clicks in Ad greater than 1000, then add the following line there:
[Measures].[clicks] > 1000
Save and close, then that filter will be valid for find elem with clicks greater than 1000.
The MDX likes below (suppose dt as dimension and clicks as measure, we want to find dt with clicks more than 1000)
WITH
SET [~ROWS] AS
Filter({[Dt].[dt].[dt].Members}, ([Measures].[clicks] > 1000))
SELECT
NON EMPTY {[Measures].[clicks]} ON COLUMNS,
NON EMPTY [~ROWS] ON ROWS
FROM [OfflineData]
i think you have two choices:
1- Add column to your fact(or view on data source view that is based on fact table)like:
case when unit_Price>2000 then 1
else 0
end as Unit_Price_Uper_Or_Under_10
and add a fictitious Dimension based on this columns value.
and add named query for New Dimension(say Range_Dimension in datasourceview :
select 1 as range
union all
select 0 as range
and after taht you cant used this filter like other dimension and attribute.
SELECT [Store].[Stores].[Store].members ON COLUMNS,
[Unit Sales] ON ROWS
FROM [Warehouse and Sales]
WHERE [Test_Dimension].[Range].&[1]
the problem is for every range you must add When condition and only if the range is static this solution is a good solution.
and for dynamic range it's better to formulate the range (based on disceretizing method )
2- add dimension with granularity near fact table based on fact table
for example if we have fact table with primary key Sale_id.we can add
dimension based on fact table with only one column sale_Id and in dimension Usage tab
we can relate this new dimension and measure group with relation type Fact and
after that in mdx we can use something like :
filter([dim Sale].[Sale Id].[Sale Id].members,[Measures].[Unit Price]>2000)