First row in group as subtotal in pivot

[michalikj]

Let's say I have a pivot table with subtotal like this (assume Expr in this example is a simple Sum, subtotals in italics):

Dim1	Dim2	Dim3	Expr
A			14
A	X		7
A	X	1	3
A	X	2	4
A	X		8
A	Y	3	6
A	Y	4	2
B			16
B	X		10
B	X	1	7
B	X	2	3
B	X		6
B	Y	3	1
B	Y	4	5

If I want to show first value of each group at the lowest grain level as the subtotal, what would be the best way to achieve this, preferably without having to touch the data model? I'm trying various combinations of Aggr() and inter-record functions (e.g. Top() ), but so far nothing works fully.

For reference, the final pivot table should look like this:

Dim1	Dim2	Dim3	Expr
A			3
A	X		3
A	X	1	3
A	X	2	4
A	X		6
A	Y	3	6
A	Y	4	2
B			7
B	X		7
B	X	1	7
B	X	2	3
B	X		1
B	Y	3	1
B	Y	4	5

Any input will be greatly appreciated.

[sunny_talwar]

You would probably need to use Dimensionality here. May be try playing around with it a little, else share a sample where you are trying to use this and we should be able to help you better

[michalikj · Accepted solution]

Nevermind I think I've got it 

For the record, formula I was looking for is this:

FirstSortedValue( 
	Aggr(
		Sum(Expr),
		(Dim1,(TEXT, ASCENDING)), (Dim2, (TEXT, ASCENDING)), (Dim3, (NUMERIC, ASCENDING))
	),
	Aggr(
		RowNo(TOTAL),
		(Dim1,(TEXT, ASCENDING)), (Dim2, (TEXT, ASCENDING)), (Dim3, (NUMERIC, ASCENDING))
	)
)

This assumes sort order in the table will remain constant, but I can live with that.