Progressive total

I think

Select m.NAME, m.date, sum(m1.Salary)
from MyTable m
LEFT OUTER JOIN MyTable m1 ON m1.date <= m.date
GROUP BY m.name, m.date

can you explain the expeted results a bit pls.   I dont understand where the figures come from

sorry, forgot a join on id.  Are the id's in your example correct, or maybe the join is supposed to be on name, not id.

Select m.NAME, m.date, sum(m1.Salary)
from MyTable m
LEFT OUTER JOIN MyTable m1 ON m1.date <= m.date AND m.id = m1.id
GROUP BY m.name, m.date

For each date that appears in the Salaries table?

BTW, avoid using reserved words like "date" are column names.
Select Name, SDate, Sum(case S.Date <= D.Date then Salary else 0 end)
FROM Salaries S CROSS JOIN Select distinct [date] as SDate from Salaries) D

Select Name, SDate, Sum(case S.Date <= D.Date then Salary else 0 end)
FROM Salaries S CROSS JOIN Select distinct [date] as SDate from Salaries) D
GROUP BY Name, SDate

Adwiseman's solution will miss out Helen's first record at least. Needs to be a cross product one the dates to ensure each date is represented for each Name.

Here with exactly your output

Select m.NAME, md.date, sum(ISNULL(m1.Salary, 0))
from (Select distinct name from MyTable) m
CROSS JOIN (Select distinct date from MyTable) md
LEFT OUTER JOIN MyTable m1 ON m1.date <= md.date AND m.name = m1.name
GROUP BY m.name, md.date
order by md.date, m.name

I'm selecting a list of names , and cross joining that with the distinct list of dates.  Goving me a list of all names, with all dates.  Then joining that in with the salart and grouping them.  The order is not the same, but the result set is.

--my working code

create table MyTable([id] int, [name] varchar(50), [date] smalldatetime, salary int)
GO
insert into MyTable
Select 1, 'Bob', '2004-01-01', 10
insert into MyTable
Select 2, 'John', '2004-01-01', 10
insert into MyTable
Select 1, 'Bob', '2004-02-01', 30
insert into MyTable
Select 3, 'Helen', '2004-02-01', 20
insert into MyTable
Select 3, 'Helen', '2004-03-01', 50
GO
select * from MyTable
GO
Select m.NAME, md.date, sum(ISNULL(m1.Salary, 0))
from (Select distinct name from MyTable) m
CROSS JOIN (Select distinct date from MyTable) md
LEFT OUTER JOIN MyTable m1 ON m1.date <= md.date AND m.name = m1.name
GROUP BY m.name, md.date
order by md.date, m.name

I would sugest using a different source for the list of all distinct dates.  A function that returns a table given a start and end date, or create a table of all dates say between 01-01-1900 and 01-01-2200

2 adwiseman  
Your code works fine
2monosodiumg
doesn't work syntax error

Thanks everybody