Combining Multiple SQL Views ON Year & Month - sql

Question

I have a SQL Server database (2012 express) with many tables.
I have produced three different VIEWS based on different combinations of the underlying tables.
Each of these views consists of three columns, Year, Month & Total
The Total column in each of the 3 Views is of a different measure.
What I want to be able to do is to combine the three Totals into a single View
I have attempted this with the following script -
SELECT b.[Year], b.[Month], b.Fees AS [Billing],
f.Estimate AS [Estimate],
w.Fees AS [WIP]
FROM MonthlyBillingTotals AS b
FULL JOIN MonthlyFeeEstimates AS f
ON (b.[Year] = f.[Year] AND b.[Month] = f.[Month])
FULL JOIN MonthlyInstructionsWIP AS w
ON (b.[Year] = w.[Year] AND b.[Month] = w.[Month])
ORDER BY b.[Year], b.[Month]
Originally I tried INNER JOINS but of course unless the Year / Month combo existed in the first view (MonthlyBillingTotals) then it did not appear in the combined query. I therefore tried FULL JOINS, but the problem here is that I get some NULLS in the Year and Month columns, when they do not exist in the first view (MonthlyBillingTotals).
If the data in the three Views is as follows -
Then what I want is -
And even better (if it is possible) -
with the missing months filled in

Answer 1

You could try building the full list of Months/Years from your tables using a UNION subquery, and then use that to drive your joins.. Something like this:
SELECT a.[Year], a.[Month], b.Fees AS [Billing],
f.Estimate AS [Estimate],
w.Fees AS [WIP]
FROM (SELECT a.[Year], a.[Month] FROM MonthlyBillingTotals AS a
UNION
SELECT b.[Year], b.[Month] FROM MonthlyFeeEstimates AS b
UNION
SELECT c.[Year], c.[Month] FROM MonthlyInstructionsWIP AS c) AS a
LEFT OUTER JOIN MonthlyBillingTotals AS b
ON (a.[Year] = b.[Year] AND a.[Month] = b.[Month])
LEFT OUTER JOIN MonthlyFeeEstimates AS f
ON (a.[Year] = f.[Year] AND a.[Month] = f.[Month])
LEFT OUTER JOIN MonthlyInstructionsWIP AS w
ON (a.[Year] = w.[Year] AND a.[Month] = w.[Month])
ORDER BY a.[Year], a.[Month]

Answer 2

This is completely untested, but see if this solves your problems:
SELECT b.[Year], b.[Month], Coalesce(b.Fees, '0') AS [Billing],
Coalesce(f.Estimate,'0') AS [Estimate],
Coalesce(w.Fees,'0') AS [WIP]
FROM MonthlyBillingTotals AS b
LEFT JOIN MonthlyFeeEstimates AS f
ON (b.[Year] = f.[Year] AND b.[Month] = f.[Month])
LEFT JOIN MonthlyInstructionsWIP AS w
ON (b.[Year] = w.[Year] AND b.[Month] = w.[Month])
ORDER BY b.[Year], b.[Month]
The Coalesce function puts in a '0' value if nothing is found, and left joins should only join parts of MonthlyFeeEstimates and MonthlyInstructionsWIP when the year and month match.

Answer 3

You could set up a small date table with year and month and left join the views with that, and use the ISNULL(variable,0) function to replace NULL with 0. Another option instead of a date table would be to use a common table expression to generate a date range to join with. In any case I suggest you look up the date table (or numbers table), it can be a really useful tool.
Edit: added an example on how a date table can be created (for reference):
declare #year_month table (y int, m int)
;with cte as (
select cast('2000-01-01' as datetime) date_value
union all
select date_value + 1
from cte
where date_value + 1 < '2010-12-31'
)
insert #year_month (y, m)
select distinct year(date_value), month(date_value)
from cte
order by 1, 2
option (maxrecursion 0)
select * from #year_month