Suppose you have a 2 dimensions cube (Units Sold) in MS Olap. Dim 1 is Product and Dim 2 is Country. In the 'Units sold' MS olap cube the cell having the coordinate ('Car', 'France') have a value (says '123').
If I set up a Pivot Table in Excel, the cell with reference ('Car', France') may show a complete different number (depending of the filters ... of course). Is there a way to have still the right number ('123') : the ideal is to have both : the filtered and the one ('123') from MS Olap.
Think 'Hide' (discrete items) instead of 'Filter' when you tick/untick items in a filter.
What you're trying to do is use one measure to do two things at the same time. Which you can't do straight off the bat. However, there are workarounds.
These are all local to Excel only (not server side) as I'm assuming it's only for you. First solution is obvious, second is better, third and fourth use an extension (fourth is my preferred method as it avoids horrendous MDX code).
1. Second Pivot Table
Create a second pivot table. First pivot table filtered by Product, Country and second pivot table by other criteria AND Product, Country. You can link the pivot tables with a slicer to choose different values for Product, Country.
2. CUBE formulas in Excel sheet cells
You don't actually need to use pivot tables to access cubes (depending on the implementation).
=CUBEVALUE("Sales","[Measures].[Profit]","[Time].[2004]","[All Product].[Beverages]")`
These CUBEVALUE formulae are based off the MDX code that goes into the Microsoft Analysis Server implementation. It goes like this:
=CUBEVALUE("CUBE NAME","[Measure Group].[Measure]","[Dimension].[Attribute]","[Dimension].[Attribute]")
You can add in whatever dimensions/attributes you want then.
3. Use the OLAP PivotTable Extension.
This allows you to create your own custom measure for individual pivot tables. Once again, using MDX. I believe you can do a similar thing like with the Cubevalue formulas - i.e. add in whatever dimensions/attributes you want on top of whatever will be filtered.
https://olappivottableextend.codeplex.com/
4. DAX Studio
DAX is the query language that powers PowerPivot. DAX studio allows you to connect to PowerPivot models and run queries - similar to SQL but just using a different language..
DAX studio allows you to connect to OLAP Cubes too. So you can perform queries against a Cube, then export it as a txt file, linked table in excel etc.
https://daxstudio.codeplex.com/
If what you'll be doing is basic, you could probably get away with 1 or 2.
However, I do highly suggest you take a look at the OLAP PivotTable Extension anyway. It has a feature which allows you to paste a list of values to filter by - instead of having to manually tick them all.
Saved me hours of woe!
Related
I want to create an empty table in a SSAS tabular cube (in SSDT, in Visual Studio) to hold all of my cube's measures. In Power BI Desktop, I would do so by creating a 1-column "empty" table in Power Query, loading it into the model, put 1 measure on it, go back into Power Query, delete the empty column, re-load the model, and continue putting measures on the table. But, I don't see an option to do this in SSDT (in Visual Studio) without using a calculated table. Is this possible? If so, how?
You don't want to do that. If the table has no columns, you won't be able to see the measures in Visual Studio:
I would create a one column table using the DAX formula ={""} and hide the column in the model and not worry about it.
If you want to name the column and have no rows, you can use this more obscure formula:
=TOPN(0, ROW("Column", ""))
Picking up (and perhaps simplifying) TheRizza's excellent idea - for which thanks - I used this:
You can obviously (I presume) create as many columns as you like. I tried putting a measure in the second column, then removing this from the table calculation formula, and SSAS moves the measure back into the first column (so you don't lose it).
Does anyone know if I can add two rows together so that I end up with just one row in Tableau (see screenshot)? So, if both rows are city Aachen and one row has a value for cost but not for purchasing power and the other row has a value for purchasing power but not cost, I would want just one row with both values. I am not interested in the columns "Table Name" and "Document Index(...". Thankful for any help!
Manipulating data like that in Tableau is usually no-go. Nevertheless, you can try Tableau prep and you should be able to do what you need here. Or maybe a different tool (even excel).
With that said, even though you have the info in two rows, the default approach for Tableau is always to aggregate data, so even if you have many rows with similar cases, once you take it to a viz using City (for example) as a dimension, this issue shouldn't really matter.
DW is written in foreign language so i translated it on pictures.
It is about some facility where you can rent rooms and equipment.
In Fact table i have one measure column which represents number of reservations and it works as expected, problem is in DimPayment table i have columns PaymentSum and PaymentSumWDiscount which i would need to use as measures. i tried to do something in calculations tab but i only get null values
I can't figure out how to use columns PaymentSum and PaymentSumWDiscount
as measures
i tried to do something in calculations tab but i only get null values
JOIN the FactReserving table to the DimPayment table in the Data Source View and then you can access those two columns in the Measure group.
It looks like you want to use your Dimension attributes as measure.
One way to achieve this is to make your Dimension table as a Measure Group as well.
The following example works with the AdventureWorks sample to create a measure group for the Product table:
In Visual Studio / SSDT, open your Cube, create a new Measure Group and select your table. (for your scenario, it would be the DimPayment table).
Your dimension usage mapping for this would be 'Fact' as both your fact and dimension are in the same table.
Now, you should be able to use the measures normally.
All the numeric fields automatically added as measures:
Using it as a measure:
I am creating an interactive 'calculator' using tableau. I have a series of dataframes that I have crossed with one another, such that the resulting dataframe is every possible combination between the tables, and every row is unique.
Each column is its own worksheet as a table. Each table in the dashboard is a pane. So, here we have a series of tables with selectable units of measurement, and the final pane on the dashboard should filter to the cell for its respective column, on the unique row of the dataset that the user has selected and 'filtered out'.
I'm having some issues getting this to work and not sure why.
The closest I can think to solving this would be 'Cascading Filters.' Here are a couple resources:
General Use
In dashboard action-filter form
The critical piece, however, is that the filters must be selected in a specific order - therefore making them 'cascading.' This may differ from your presumed concept of clicking/filtering in any order on the worksheets to then arrive to a final answer. I do think that this may be a limitation of Tableau - I don't think that a 'many to many' type of relationship can be set up within Action Filters.
Is it possible to create a table with this crystal reports, because I need to layout or format my data into a table(not cross tab), is there a way to do this cause I think my tool box is only limited to lines and boxes and basic shapes.
There's a few things you can do to make it look like a table:
Create Each row in your report represents a row in the table.
Put a border/box around the table use the 'Insert Box' from the toolbar and simple start drawing the box in the header of the (or group header) and finish in the footer (or group footer) and when you preview the report you'll see a box around the whole table [report]
Column grid lines can easily be done by linking the top and bottom of your box. they will then expand out in preview mode.
Top Tip: Zooming right in, to say 300%, makes getting the joins much easier, and investigate snap-to-grid you may want to simple disable it it may actually make life easier,(when ultimate precision is required)
If you need to display the data as a table, you can get creative with details, groups, subsections, and formulae. For example, the formula sum({A},{B}) adds all the values of {A} for each group {B} (best used in a group footer). If you have any specific questions, I can try to answer them.
Sorry, but Crystal reports can not create tables. It can execute a stored procedure that creates tables, but that's really the procedure creating the tables.