Counting number of rows grouped by date and hour - sql

Question

I am tracking customer store entry data in Microsoft SQL Server 2008 R2 that looks something like this:
DoorID DateTimeStamp EntryType
1 2013-09-02 09:01:16.000 IN
1 2013-09-02 09:04:09.000 IN
1 2013-09-02 10:19:29.000 IN
1 2013-09-02 10:19:30.000 IN
1 2013-09-02 10:19:32.000 OUT
1 2013-09-02 10:26:36.000 IN
1 2013-09-02 10:26:40.000 OUT
I don't want to count the OUT rows, just IN.
I believe that it needs to be grouped on Date, and DoorID, then get the hours totals.
I would like it to come out like this.
Date DoorID HourOfDay TotalInPersons
2013-09-02 1 0 0
2013-09-02 1 1 0
2013-09-02 1 2 0
2013-09-02 1 3 0
2013-09-02 1 4 0
2013-09-02 1 5 0
2013-09-02 1 6 0
2013-09-02 1 7 0
2013-09-02 1 8 0
2013-09-02 1 9 2
2013-09-02 1 10 3
2013-09-02 1 11 0
2013-09-02 1 12 0
2013-09-02 1 13 0
2013-09-02 1 14 0
2013-09-02 1 15 0
2013-09-02 1 16 0
2013-09-02 1 17 0
2013-09-02 1 18 0
2013-09-02 1 19 0
2013-09-02 1 20 0
2013-09-02 1 21 0
2013-09-02 1 22 0
2013-09-02 1 23 0

Answer 1

SELECT
[Date] = CONVERT(DATE, DateTimeStamp),
DoorID,
HourOfDay = DATEPART(HOUR, DateTimeStamp),
TotalInPersons = COUNT(*)
FROM dbo.tablename
WHERE EntryType = 'IN'
GROUP BY
CONVERT(DATE, DateTimeStamp),
DoorID,
DATEPART(HOUR, DateTimeStamp)
ORDER BY
[Date], DoorID, HourOfDay;
Of course if you need all hours, even where no rows are represented, here is one solution (which limits the output for any day only to the doors that have at least one IN entry on that day):
;WITH h AS
(
SELECT TOP (24) h = number FROM Master..spt_values
WHERE type = N'P' ORDER BY number
),
doors AS
(
SELECT DISTINCT DoorID, [Date] = CONVERT(DATE,DateTimeStamp)
FROM dbo.tablename WHERE EntryType = 'IN'
)
SELECT
d.[Date],
d.DoorID,
HourOfDay = h.h,
TotalInPersons = COUNT(t.EntryType)
FROM doors AS d CROSS JOIN h
LEFT OUTER JOIN dbo.tablename AS t
ON CONVERT(DATE, t.DateTimeStamp) = d.[Date]
AND t.DoorID = d.DoorID
AND DATEPART(HOUR, t.DateTimeStamp) = h.h
AND t.EntryType = 'IN'
GROUP BY d.[Date], d.DoorID, h.h
ORDER BY d.[Date], d.DoorID, h.h;

Answer 2

How about something like this:
SELECT
CAST(DateTimeStamp AS DATE) AS Date
,DoorID
,DATEPART(HOUR, DateTimeStamp) AS HourOfDay
,COUNT(*) AS TotalInPersons
FROM StoreTable
WHERE EntryType = 'IN'
GROUP BY
CAST(DateTimeStamp AS DATE)
,DoorID
,DATEPART(HOUR, DateTimeStamp)

Answer 3

This should work. I guessed on how you would pull DoorID and TotalPersons, but the overall logic is correct
SELECT CONVERT(date,dateColumn) AS Date,
datepart(hh,dateColumn) AS HourOfDay,
DoorID,
COUNT(people) AS TotalPersons
FROM yourtable
WHERE EntryType = 'IN'
GROUP BY CONVERT(date,dateColumn), datepart(hh,dateColumn), DoorID
ORDER BY CONVERT(date,dateColumn), datepart(hh,dateColumn)