Reputation: 113
I need to query a sales table and have the resulting table show the breakdown for different time periods within the given day.
For example, the table has the following fields:
Id (int), EntryDate (datetime), SaleReference (varchar)
I'd like to produce a table that looks like this:
Date Sales 9am-12pm Sales 12pm-3pm Sales 6pm-9pm ---------- -------------- -------------- ------------- 01-01-2010 10 20 6 02-01-2010 12 16 3 03-01-2010 43 11 2
Any help on this would be greatly appreciated. Thanks
Upvotes: 5
Views: 5812
Reputation: 453287
Assuming SQL Server below. If not a similar logic will probably apply with your RDBMS but likely a different function to get the hour part from the datetime and the behaviour of BETWEEN may be different too (In SQL Server it is an inclusive range).
SELECT CAST([Date] AS Date) AS [Date],
COUNT(CASE WHEN DATEPART(hour, [Date]) BETWEEN 9 AND 11 THEN 1 ELSE NULL END)
AS [Sales 9am-12pm],
COUNT(CASE WHEN DATEPART(hour, [Date]) BETWEEN 12 AND 14 THEN 1 ELSE NULL END)
AS [Sales 12pm-3pm],
COUNT(CASE WHEN DATEPART(hour, [Date]) BETWEEN 18 AND 20 THEN 1 ELSE NULL END)
AS [Sales 6pm-9pm]
FROM Table
GROUP BY CAST([Date] AS Date) /*For SQL2008*/
NB: Previous versions of SQL Server require a few more hoops to get just the date part out of a datetime. e.g. CAST(FLOOR( CAST( GETDATE() AS FLOAT ) ) AS DATETIME
(From here)
Upvotes: 5
Reputation: 2694
Assuming your database supports this kind of date math, you can say:
CASE WHEN EntryDate - date(EntryDate) >= INTERVAL '9 hours'
AND EntryDate - date(EntryDate) < INTERVAL '12 hours'
THEN ...
(That's the PostgreSQL interval syntax, btw... might be nonstandard.) But there are probably more elegant ways of doing it.
Upvotes: 1