Question

Possible Duplicate:
SQL Server: How to select all days in a date range even if no data exists for some days

I wasn’t really sure how to word this question, but I’ll try to explain. I’m trying to build some basic reporting using queries like the following:

SELECT COUNT(*) AS count, h_date FROM (SELECT CONVERT(VARCHAR(10), h_time, 102) AS h_date FROM hits h GROUP BY h_date ORDER BY h_date

This returns results like this, which I use to build a graph:

8 2012.05.06
2 2012.05.07
9 2012.05.09

As you can see, it’s missing the 8th as there were no hits on that day. Is there a way to get a value of 0 for dates that have no results, or will I have to parse the results after the fact and add them manually?

Answer 1

You can use existing catalog views to derive a sequential range of dates between your start date and your end date. Then you can just left join to your data, and any missing dates will be there with 0s.

DECLARE @min SMALLDATETIME, @max SMALLDATETIME;

SELECT @min = MIN(h_time), @max = MAX(h_time)
  FROM dbo.hits 
  -- WHERE ?

-- or if you just want a fixed range:

-- SELECT @min = '20120101', @max = '20120131';

;WITH n(d) AS 
(
  SELECT TOP (DATEDIFF(DAY, @min, @max)+1) 
   DATEADD(DAY, ROW_NUMBER() OVER (ORDER BY [object_id]) - 1, DATEDIFF(DAY, 0, @min))
  FROM sys.all_objects ORDER BY [object_id]
)
SELECT n.d, [count] = COUNT(h.h_time)
  FROM n
  LEFT OUTER JOIN dbo.hits AS h
  ON h.h_time >= n.d
  AND h.h_time < DATEADD(DAY, 1, n.d)
  -- AND --WHERE clause against hits?
  GROUP BY n.d;