How to sum only the first row for each group in a result set - sql

Question

Ok, I will try to explain myself the best I can, but I have the following:
I have a datasource that basically is a dynamic query. The query in itself shows 3 fields, Name, Amount1, Amount2.
Now, I could have rows with the same Name. The idea is to make a sum of Amount1+Amount2 WHEN Name is distinct from the previous one I saved. If I would do this on C# it could be something like this:
foreach (DataRow dr in repDset.Dataset.Rows)
{
total = (long)dr["Amount1"] + (long)dr["Amount2"];
if (thisconditiontrue)
{
if (PreviousName == "" || PreviousName != dr["Name"].ToString())
{
TotalName = TotalName + total;
}
PreviousName = dr["Name"].ToString();
}
}
The idea is to grab this and make a Reporting Services expression using the methods RS can give me, for example:
IIF(Fields!Name.Value<>Previous(Fields!Name.Value),Fields!Amount1.Value + Fields!Amount2.Value,False)
Something like that but that stores the amount of the previous one.
Maybe creating another field? a calculated one?
I can clarify further and edit if needed.
*EDIT for visual clarification:
As an example, it is something like this:

Answer 1

This query is assuming you're working with SQL server. But you're going to need something to order the query results by otherwise how do you know which row is the first one?
SELECT SUM(NameTotal) AS Total
FROM (
SELECT Name, Amount1 + Amount2 AS NameTotal,
ROW_NUMBER() OVER (ORDER BY OrderField PARTITION BY Name) AS rowNum
FROM srcTable
) AS a
WHERE rowNum=1;
This uses the analytical window function ROW_NUMBER() to number each row and the PARTITION BY clause tells it to reset the numbering for every different value of Name in the result set. You do need a field that you can order the results by though or this won't work. If you really just want a random order you can do ORDER BY NEWID() but that will give you a non-deterministic result.
This syntax is particular to SQL server but it can usually be achieved in other databases.
If you're looking to display the output like you've shown in your example you could use two queries and reference the other one by passing it as the scope to an aggregate function in an SSRS expression like this:
=MAX(Fields!Total.Value, "TotalQueryDataset")
Where your dataset is called "TotalQueryDataset".
Otherwise you can achieve the output using pure SQL like this:
WITH nameTotals AS (
SELECT Name, Amount1, Amount2,
ROW_NUMBER() OVER (ORDER BY OrderField PARTITION BY Name) AS rowNum
FROM srcTable
)
SELECT Name, Amount1, Amount2
FROM nameTotals
UNION ALL
SELECT 'Total', SUM(Amount1 + Amount2), NULL
FROM nameTotals
WHERE rowNum=1;