Reputation: 57
I've the following Datatable:
[Date];[Time];[ProcessID];[ID]
24.09.2012;08:18:00;4000;1
24.09.2012;08:18:00;4000;2
24.09.2012;08:19:00;4001;3
24.09.2012;08:23:00;4002;4
24.09.2012;08:32:00;4003;5
...
As result I would like to have a grouping by a 15-minute Time-Interval to count the ProcessID´s:
[Interval];[ProcessID Count]
8:15:00 - 8:29:00;3
8:30:00 - 8:45:00;1
How to achieve this goal?
Upvotes: 3
Views: 1281
Reputation: 10931
Here's the solution against Linq-to-SQL (on SQL Server) with the DateTime being in a single field (a LINQpad query):
Const BaseDate As Date = #9/24/2012#
Const PeriodMinutes As Integer = 15
Dim p = From t In TestData _
Group By q=CInt(Math.Floor((t.SomeDate - BaseDate).TotalMinutes/PeriodMinutes)) Into Group _
Select s=BaseDate.AddMinutes(q*PeriodMinutes), _
e=BaseDate.AddMinutes((q+1)*PeriodMinutes), _
ids=Aggregate g In Group Select g.ProcessID Distinct Into Count()
p.Dump
Using different types and other "little" adjustments could produce simpler SQL backend code.
For your actual database, you need to adjust the Group By
, although it now depends upon what your Time column is converted to through Linq-to-SQL and/or your database's driver.
As it is a TimeSpan
the Group By
becomes something like:
Group By d=t.[Date], q=CInt(Math.Floor(t.[Time].TotalMinutes/PeriodMinutes)) Into Group _
Select s=d.AddMinutes(q*PeriodMinutes), _
e=d.AddMinutes((q+1)*PeriodMinutes), _
ids=Aggregate g In Group Select g.ProcessID Distinct Into Count()
Upvotes: 1
Reputation: 460048
This should work (however, not tested):
Dim groupedByQuarter =
From r In table
Let day = r.Field(Of Date)("Date").Date
Let time = r.Field(Of TimeSpan)("Time")
Let quarter = time.Add(TimeSpan.FromMinutes((-1) * time.Minutes Mod 15))
Group r By Key = New With {Key .day = day, Key .quarter = quarter} Into TimeGroup = Group
Select New With {
.Interval = String.Format("{0} - {1}",
Key.quarter,
Key.quarter.Add(TimeSpan.FromMinutes(14))),
.ProcessCount = TimeGroup.Select(Function(r) r.Field(Of Int32)("ProcessID")).Distinct().Count()
}
Upvotes: 1