Creating a "pivot table" like construct in sql - sql

Question

This is going to be a bit convoluted but I have a table named dbo.sizeclassreport. I apologize if this is a bit vague.
It has 78656 rows/records which is not all that immense. There are three salient fields; employment, area, and sizeclass. Sample data below
Area Employment sizeclass
000003 4 01
000001 6 02
000013 12 03
000003 2 01
000005 23 04
000001 7 02
000003 11 03
From here, the result would look this. There are 9 sizeclasses and 18 areas. Need to produce 9 breakdowns of each area. Each area would have a breakdown by size class.
Area Sizeclass Employment in Size Class Number of Sites
Total 00 17(6+11) 3
000003 01 6 (4+2) 2
000003 02 NULL or Empty NULL or Empty
000003 03 11 1
......
000003 09 NULL or Empty NULL or Empty

Answer 1

SELECT
[Areas].[Area],
[classSizes].[sizeclass],
SUM([data].[Employment]) AS [Employment in Size Class],
/*this would be better with your PK */
CASE COUNT([data].[Employment])
WHEN 0 THEN NULL
ELSE COUNT([data].[Employment])
END AS [Number of Sites]
/*These rows create your base matix*/
FROM (SELECT DISTINCT
[Area]
FROM [sizeclassreport]) AS [Areas]
CROSS JOIN (SELECT DISTINCT
[sizeclass]
FROM [sizeclassreport]) AS [classSizes]
/* this is where you select your data*/
LEFT OUTER JOIN [sizeclassreport] [data]
ON [Areas].[Area] = [data].[Area]
AND [classSizes].[sizeclass] = [data].[sizeclass]
GROUP BY
[areas].[Area],
[classSizes].[sizeclass]
ORDER BY
[areas].[Area],
[classSizes].[sizeclass]

Answer 2

It looks like you want all areas and sizeclasss in the output. If so, use cross join to generate the results, then left join and group by for the calculations:
select a.area, sc.sizeclass,
sum(scr.employment) as employment,
count(scr.area) as number_of_sizes
from (select distinct area from dbo.sizeclassreport scr) a cross join
(select distinct sizeclass from dbo.sizeclassreport scr) sc left join
dbo.sizeclassreport scr
on a.area = scr.area and sc.sizeclass = scr.sizeclass
group by a.area, sc.sizeclass;
Note: This returns 0 instead of NULL for the last column. This makes more sense to me. However, you could use NULLIF() if you really want a NULL value.