How to calculate running total (month to date) in SQL Server 2008 - sql

Question

I'm trying to calculate a month-to-date total using SQL Server 2008.
I'm trying to generate a month-to-date count at the level of activities and representatives. Here are the results I want to generate:
| REPRESENTATIVE_ID | MONTH | WEEK | TOTAL_WEEK_ACTIVITY_COUNT | MONTH_TO_DATE_ACTIVITIES_COUNT |
|-------------------|-------|------|---------------------------|--------------------------------|
| 40 | 7 | 7/08 | 1 | 1 |
| 40 | 8 | 8/09 | 1 | 1 |
| 40 | 8 | 8/10 | 1 | 2 |
| 41 | 7 | 7/08 | 2 | 2 |
| 41 | 8 | 8/08 | 4 | 4 |
| 41 | 8 | 8/09 | 3 | 7 |
| 41 | 8 | 8/10 | 1 | 8 |
From the following tables:
ACTIVITIES_FACT table
+-------------------+------+-----------+
| Representative_ID | Date | Activity |
+-------------------+------+-----------+
| 41 | 8/03 | Call |
| 41 | 8/04 | Call |
| 41 | 8/05 | Call |
+-------------------+------+-----------+
LU_TIME table
+-------+-----------------+--------+
| Month | Date | Week |
+-------+-----------------+--------+
| 8 | 8/01 | 8/08 |
| 8 | 8/02 | 8/08 |
| 8 | 8/03 | 8/08 |
| 8 | 8/04 | 8/08 |
| 8 | 8/05 | 8/08 |
+-------+-----------------+--------+
I'm not sure how to do this: I keep running into problems with multiple-counting or aggregations not being allowed in subqueries.

Answer 1

A running total is the summation of a sequence of numbers which is
updated each time a new number is added to the sequence, simply by
adding the value of the new number to the running total.
I THINK He wants a running total for Month by each Representative_Id, so a simple group by week isn't enough. He probably wants his Month_To_Date_Activities_Count to be updated at the end of every week.
This query gives a running total (month to end-of-week date) ordered by Representative_Id, Week
SELECT a.Representative_ID, l.month, l.Week, Count(*) AS Total_Week_Activity_Count
,(SELECT count(*)
FROM ACTIVITIES_FACT a2
INNER JOIN LU_TIME l2 ON a2.Date = l2.Date
AND a.Representative_ID = a2.Representative_ID
WHERE l2.week <= l.week
AND l2.month = l.month) Month_To_Date_Activities_Count
FROM ACTIVITIES_FACT a
INNER JOIN LU_TIME l ON a.Date = l.Date
GROUP BY a.Representative_ID, l.Week, l.month
ORDER BY a.Representative_ID, l.Week
| REPRESENTATIVE_ID | MONTH | WEEK | TOTAL_WEEK_ACTIVITY_COUNT | MONTH_TO_DATE_ACTIVITIES_COUNT |
|-------------------|-------|------|---------------------------|--------------------------------|
| 40 | 7 | 7/08 | 1 | 1 |
| 40 | 8 | 8/09 | 1 | 1 |
| 40 | 8 | 8/10 | 1 | 2 |
| 41 | 7 | 7/08 | 2 | 2 |
| 41 | 8 | 8/08 | 4 | 4 |
| 41 | 8 | 8/09 | 3 | 7 |
| 41 | 8 | 8/10 | 1 | 8 |
SQL Fiddle Sample

Answer 2

As I understand your question:
SELECT af.Representative_ID
, lt.Week
, COUNT(af.Activity) AS Qnt
FROM ACTIVITIES_FACT af
INNER JOIN LU_TIME lt ON lt.Date = af.date
GROUP BY af.Representative_ID, lt.Week
SqlFiddle

Answer 3

Representative_ID Week Month_To_Date_Activities_Count
41 2013-08-01 00:00:00.000 1
41 2013-08-08 00:00:00.000 3
USE tempdb;
GO
IF OBJECT_ID('#ACTIVITIES_FACT','U') IS NOT NULL DROP TABLE #ACTIVITIES_FACT;
CREATE TABLE #ACTIVITIES_FACT
(
Representative_ID INT NOT NULL
,Date DATETIME NULL
, Activity VARCHAR(500) NULL
)
IF OBJECT_ID('#LU_TIME','U') IS NOT NULL DROP TABLE #LU_TIME;
CREATE TABLE #LU_TIME
(
Month INT
,Date DATETIME
,Week DATETIME
)
INSERT INTO #ACTIVITIES_FACT(Representative_ID,Date,Activity)
VALUES
(41,'7/31/2013','Chat')
,(41,'8/03/2013','Call')
,(41,'8/04/2013','Call')
,(41,'8/05/2013','Call')
INSERT INTO #LU_TIME(Month,Date,Week)
VALUES
(8,'7/31/2013','8/01/2013')
,(8,'8/01/2013','8/08/2013')
,(8,'8/02/2013','8/08/2013')
,(8,'8/03/2013','8/08/2013')
,(8,'8/04/2013','8/08/2013')
,(8,'8/05/2013','8/08/2013')
--Begin Query
SELECT AF.Representative_ID
,LU.Week
,COUNT(*) AS Month_To_Date_Activities_Count
FROM #ACTIVITIES_FACT AS AF
INNER JOIN #LU_TIME AS LU
ON AF.Date = LU.Date
Group By AF.Representative_ID
,LU.Week