sql query to find sum of all rows and count of duplicates

Question

If data is in the following format:

SID  TID  Tdatetime        QID   QTotal  
----------------------------------------
100  1    01/12/97 9:00AM  66    110   
100  1    01/12/97 9:00AM  66    110  
100  1    01/12/97 10:00AM 67    110  
100  2    01/19/97 9:00AM  66    .  
100  2    01/19/97 9:00AM  66    110  
100  2    01/19/97 10:00AM 66    110  
100  3    01/26/97 9:00AM  68    120  
100  3    01/26/97 9:00AM  68    120  
110  1    02/03/97 10:00AM 68    110  
110  3    02/12/97 9:00AM  64    115  
110  3    02/12/97 9:00AM  64    115  
120  1    04/05/97 9:00AM  66    105  
120  1    04/05/97 10:00AM 66    105  

I would like to be able to write a query to sum the QTotal column for all rows and find the count of duplicate rows for the Tdatetime column.

The output would look like:

  Year   Total  Count

     97 | 1340 | 4

The third column in the result does not include the count of distinct rows in the table. And the output is grouped by the year in the TDateTime column.

Answer 1

This will work if you really want to group by the tDateTime column:

SELECT DISTINCT tDateTime, SUM(QTotal), Count(distinct tDateTime)
FROM Table
GROUP BY tDateTime
HAVING  Count(distinct tDateTime) > 1

But your results look like you want to group by the Year in the tDateTime column. Is this correct?

If so try this:

SELECT DISTINCT YEAR (tDateTime), SUM(QTotal), Count(distinct tDateTime)
FROM Table
GROUP BY YEAR (tDateTime)
HAVING  Count(distinct tDateTime) > 1

Answer 2

The following query may help:

SELECT 
    'YEAR ' + CAST(sub.theYear AS VARCHAR(4)), 
    COUNT(sub.C), 
    (SELECT SUM(QTotal) FROM MyTable WHERE YEAR(Tdatetime) = sub.theYear) AS total
FROM 
   (SELECT 
        YEAR(Tdatetime) AS theYear, 
        COUNT(Tdatetime) AS C 
    FROM MyTable 
    GROUP BY Tdatetime, YEAR(Tdatetime)
    HAVING COUNT(Tdatetime) >= 2) AS sub

Answer 3

You must do SELECT from this table GROUPing by QTotal, using COUNT(subSELECT from this table WHERE QTotal is the same). If I only I had time I would write you SQL statement, but it'll take some minutes.

Answer 4

Something like:

select Year(Tdatetime)  ,sum(QTotal), count(1) from table group by year(Tdatetime )

or full date

select Tdatetime  ,sum(QTotal), count(1) from table group by year(Tdatetime) 

Or your ugly syntax ( : ) )

select 'Year ' + cast(Year(tdatetime) as varchar(4)) 
     + '|' + cast(sum(QTotal) as varchar(31)) 
     + '|' + cast(count(1) as varchar(31)) 
 from table group by year(Tdatetime )

Or do you want just the year? Sum all columns? Or just by year?

Answer 5

SELECT
 YEar + year(Tdatetime),
 SUM ( QTotal ),
 (SELECT COUNT(*) FROM ( 
 SELECT Tdatetime FROM tDateTime GROUP BY Tdatetime        
 HAVING COUNT(QID) > 1) C
FROM
 Tdatetime t

GROUP BY 
 YEar + year(Tdatetime)

Answer 6

This is the first time I have asked a question on stackoverflow. It looks like I have lost my original ID info. I had to register to login and add comments to the question I posted.

To answer OMG Ponies question, this is a SQL Server 2008 database. @Abe Miessler , the row with SID 120 does not contain duplicates. the first row for SID 120 shows 9:00AM in the datetime column , and the second row shows 10:00AM.

@Zafer, your query is the accepted answer. I made a few minor tweaks to get it to work. Thanks. Thanks due to Abe Miessler and the others for your help.