Question

I have the code below created that provides me with a count of unique users broken down by company. I would like to add a date modifier that would have two columns, one for “Last Week” and one for “2 Weeks Ago”. I would like these columns to respectively show active users for the previous week’s dates and the week prior to that based off the field “users.date”. Any help would be fantastic. I am using SQL Server Management Studio 2008 R2.

 --Active Users
    SELECT company.companyName as 'Group Name',
    COUNT(distinct users.userid) [Count]
    FROM Users, Company 
    WHERE 
    Jobstate = '6'
    and company.companyID = users.companyid
    and company.companyID in (1,4,31)
    GROUP BY company.companyName

Currently I am receiving this as a result:

Group Name | Count     |
------------------------
Company 1  | 104       |
Company 2  | 74        |
Company 3  | 46        |

What I would like to see would be:

Group Name | Last Week | 2 Weeks Ago |
--------------------------------------
Company 1  | 14        | 16          |
Company 2  | 7         | 12          |
Company 3  | 4         | 8           |

Answer 1

For SQL Server it would look something like this.

SELECT
  Company.CompanyName AS [Group Name],
  COUNT(DISTINCT Users.UserId) AS [Count],
  SUM (CASE 
         WHEN DATEDIFF(day, Users.Date, GETDATE()) < 7 THEN 1
         ELSE 0
       END
      ) AS [Last Week]
  -- Additional columns for last 2 weeks, etc.
FROM Users, Company 

WHERE 
  Jobstate = '6'
  AND Company.CompanyId = Users.CompanyId
  AND Company.CompanyId in (1, 4, 31)
GROUP BY
  Company.CompanyName

EDIT: The query above counts users, not distinct users. If you want distinct users for the last week it can be done very simply by modifying your original query.

SELECT
  Company.CompanyName AS [Group Name],
  COUNT(DISTINCT Users.UserId) AS [Last Week]
FROM Users, Company 

WHERE 
  Jobstate = '6'
  AND Company.CompanyId = Users.CompanyId
  AND Company.CompanyId in (1, 4, 31)
  AND DATEDIFF(day, Users.Date, GETDATE()) < 7
GROUP BY
  Company.CompanyName

If you want to combine the two so you have all the columns in a single dataset it can be done by combining the two approaches like this:

SELECT
  Company.CompanyName AS [Group Name],
  COUNT(DISTINCT Users.UserId) AS [Count],
  COUNT (DISTINCT
           CASE 
             WHEN DATEDIFF(day, Users.Date, GETDATE()) < 7 THEN Users.UserId
             ELSE NULL
           END
        ) AS [Last Week]
  -- Additional columns for last 2 weeks, etc.
FROM Users, Company 

WHERE 
  Jobstate = '6'
  AND Company.CompanyId = Users.CompanyId
  AND Company.CompanyId in (1, 4, 31)
GROUP BY
  Company.CompanyName

SQLFiddle here