Question

I am running the script below. I was wondering if someone knows how to add three additional rows to the result:

1.add a new row with the MAX value

2.add a new row with the MIN value.

3.add a new row with the average.

Here is my query which run on two full month of April and May of 2006 data.

SELECT
DATEPART(YYYY, DATEADD(MM, 3, SOLD_DATE)) AS FY,
DATENAME(WEEKDAY, SOLD_DATE)              AS DAY,
STORE_NAME                                AS STORE,
CONVERT (VARCHAR, SOLD_DATE, 10)          DATES,
SUM(ITEMS)                                AS ITEM,
'NUMBER'                                  AS NOTE     
FROM MYTABLE
WHERE SOLD_DATE >='04/1/2006'
AND SOLD_DATE <'06/1/2006'
AND STORE_NAME ='ELEVEN'
GROUP BY DATEPART(YYYY, DATEADD(MM, 3, SOLD_DATE)), 
DATENAME(WEEKDAY, SOLD_DATE), STORE_NAME, SOLD_DATE

The query yields the following result(partial), with the desired additional rows at the bottom of the result.

FY     DAY        STORE   DATES     ITEM    NOTE
2006    Saturday    ELEVEN  4/1/2006    14  NUMBER
2006    Sunday      ELEVEN  4/2/2006    21  NUMBER
2006    Monday      ELEVEN  4/3/2006    24  NUMBER
--------------------------------------------------
--------------------------------------------------
--------------------------------------------------
2006    Monday      ELEVEN  5/29/2006   37  NUMBER
2006    Tuesday     ELEVEN  5/30/2006   20  NUMBER
2006    Wednesday   ELEVEN  5/31/2006   25  NUMBER
2006    Saturday    ELEVEN  5/13/2006   5   MINIMUM
2006    Tuesday     ELEVEN  5/16/2006   61  MAXIMUM
2006                ELEVEN              25  AVERAGE

Answer 1

Suggest you load your query into a table variable first. Then you can perform your meta operations. Then UNIONing the 3 rows onto the resultset.

The benefit of this approach:

• you’re calculating from the SAME resultset. If you copy/pasted the query for the 3 (min,max,avg), you potentially could be reading different data, if there’s another process who’d updated the rows while you were calculating. If/when this happens, your calculations would appear to be incorrect, compared to the raw data in the resultset.

• much less code to read and maintain

DECLARE @MyResults (TABLE) (FY int, DAY varchar(10), STORE varchar(10),   DATES varchar(10), ITEM varchar(10),    NOTE varchar(10))

DECLARE   @MyCalcs TABLE (ForYear int, Item int, Note varchar(100))

--calc the avg, max and min

INSERT INTO @MyCalcs (ForYear, Item, Note)
   SELECT  FY, AVG(ITEMS), 'Average'
   FROM @MyResults GROUP BY FY

INSERT INTO @MyCalcs (ForYear, Item, Note)
   SELECT  FY, Max(ITEMS), 'Max'
   FROM @MyResults GROUP BY FY

INSERT INTO @MyCalcs (ForYear, Item, Note)
   SELECT  FY, MIN(ITEMS), 'Min'
   FROM @MyResults GROUP BY FY

--union the discrete data with the calcs
SELECT FY,DAY,STORE,DATES,ITEM,NOTE
FROM @MyResults
UNION ALL
SELECT ForYear AS FY, '' AS Day, '' AS Store, Item, Note
FROM @MyCalcs