SQL Summing Multiple Joins

Question

I shortened the code quite a bit, but hopefully someone will get the idea of what i am tryign to do. Need to sum totals from two different selects, i tried putting each of them in Left Outer Joins(tried Inner Joins too). If i run wiht either Left Outer Join commented out, I get the correct data, but when i run them together, i get really screwed up counts. So, i know joins are probably not the correct approach to summing data from the same table, i can;t simple do it in a where clause there is other table involved int he code i commented out.

I guess i am trying to sum together 2 different queries.

SELECT  eeoc.EEOCode AS 'Test1',
        SUM(eeosum.Col_One) AS 'Col_One',

FROM EEO1Analysis eeo
    LEFT OUTER JOIN (
        SELECT eeor.AnalysisID, eeor.Test1,
                SUM(CASE eeor.ZZZ WHEN 1 THEN (CASE eeor.AAAA WHEN 1 THEN 1 ELSE 0 END) ELSE 0 END) AS 'Col_One',
        FROM EEO1Roster eeor
            ..........
        WHERE  eeor.AnalysisID = 7 
        GROUP BY eeor.AnalysisID, eeor.EEOCode
    ) AS eeosum2 ON eeosum2.AnalysisID = eeo.AnalysisID
    LEFT OUTER JOIN (
        SELECT eeor.AnalysisID, eeor.Test1,
                SUM(CASE eeor.ZZZ WHEN 1 THEN (CASE eeor.AAAA WHEN 1 THEN 1 ELSE 0 END) ELSE 0 END) AS 'Col_One',
        FROM EEO1Roster eeor
            ........
    ) AS eeosum ON eeosum.AnalysisID = eeo.AnalysisID 

WHERE  eeo.AnalysisID = 7
GROUP BY eeoc.Test1

Answer 1

What about

SELECT ... FROM EEO1Analysis eeo,

  (SELECT ... LEFT OUTER JOIN ... GROUP BY ... ) AS data

...     

?

And, if you can, I'd recommend preparing the data to separate tables, then operate on them with different analysis IDs. Could save some execution time at least.

Answer 2

Need to sum totals from two different selects If you expect one row single-column result, this way is enough

SELECT 
   ((SELECT SUM(...) FROM ... GROUP BY...) + 
   (SELECT SUM(...) FROM ... GROUP BY...)) as TheSumOfTwoSums

Answer 3

You could UNION ALL the 2 queries and then do a SUM + GROUP BY i.e.

SELECT Col1, Col2, SUM(Col_One) FROM
     (SELECT Col1, Col2, SUM(Col_One) 
          FROM Table1 
          WHERE <Conditionset1> 
          GROUP BY Col1, Col2
     UNION ALL
     SELECT Col1, Col2, SUM(Col_One)
          FROM Table1 
          WHERE <Conditionset2> 
          GROUP BY Col1, Col2)
GROUP BY
     Col1, Col2

Of course, if there is are row(s) returned by and they would be double counted.