Help needed - How to create a single row table with the value from a selectedvalue.

[andreyinfo]

I need to create a table and then with this table, to create another table. But the first table need to be created from a selection.

Tried everything and I just can't go on.
The measure getting the selectedvalue:

Selected = SELECTEDVALUE(dProducts[ProductCode])

The first table attempt :

FilteredBySelectedValue = 
Var _ProductCode = [Selected]
Return
ROW( "ProductCode",_ProductCode)

The second table attempt :

fBasketRelated = 
FILTER(
CROSSJOIN(
VALUES(FilteredBySelectedValue[ProductCode]),SELECTCOLUMNS( VALUES( fSalesDetails[ProductCode]),"RProductCode",[ProductCode])),[ProductCode]>[RProductCode])

The result : Nowhere.

f1

The model :

f2
@stevedep 

 The PBI file 

[lbendlin]

I think it depends what you expect to do with the result.  If you start constructing a table from a selectedvalue in a filter context then you can use that table in memory and transform it as needed but the final return value needs to be scalar (a count of rows, or a CONCATENATEX of values).

[andreyinfo]

So, if got it right, my final result needs to be a measure, because a table return more than one result.

If so, there is no solution for my needs?

Because I need to create a table : 

Table = 
FILTER(
    CROSSJOIN(
        VALUES(FilteredBySelectedValue[ProductCode]),SELECTCOLUMNS( VALUES( fSalesDetails[ProductCode]),"RProductCode",[ProductCode])),[ProductCode]>[RProductCode])

 And the first values ("VALUES(FilteredBySelectedValue[ProductCode])") needs to come from a selectedvalue.

[lbendlin]

your final result type must match the type of the original value.  

How do you expect to consume the table that you created out of  a scalar value?

[andreyinfo]

I thought that with this code I would create a table with a single row and column(Figure 1), and it seems a scalar result because it is only one of it's kind. Please, note that if I set the value manually (VAR _ProductCodeTyped ) I get the desired value and I can consume the table to generate the second table (figure 2).

Figure 1

And then with the second table I would create columns and measures.

Figure 2

If I got your question right. @lbendlin .

[Anonymous · Accepted solution]

That's not possible. Tables in a model are always static and loaded at load time (then set in stone!).

Secondly, you can only create table variables in dynamic code that describes measures but such tables are gone as soon as the measure's been calculated.

Best
D

[lbendlin]

you might get away with a single row/single column table as the final output. But not as an intermediate step to a larger table.