LINQ, Join, GroupBy and Count - vb.net

Question

I have this vb.net LINQ query. Without the join the count and SUM worked perfectly on each. But with the join the count and SUM arent working properly due to the multiple CAMPAIGNID values. I suspect I need another group by statement in there but I am not entirely sure how to do that. IF you could help that would be great.
Dim query =
From t1 In tbl1
Join t2 In tbl2 On t1.CAMPAIGNID Equals t2.CAMPAIGNID
Group By t1.CAMPAIGNID Into Group
Select New With {
.id = CAMPAIGNID,
.CALLS = Group.Sum(Function(a) a.t2.CALLS),
.count = Group.Count(Function(a) a.t1.TERMCD = "Refused")
}

Answer 1

I have made an example which seems to be working properly.
Public Class T1
Public Property CAMPAIGNID As Integer
Public Property TERMCD As String
End Class
Public Class T2
Public Property CAMPAIGNID As Integer
Public Property CALLS As Integer
End Class
Dim tbl1 As New List(Of T1) From {
New T1 With {.CAMPAIGNID = 1, .TERMCD = "Refused"},
New T1 With {.CAMPAIGNID = 1, .TERMCD = "Accepted"},
New T1 With {.CAMPAIGNID = 2, .TERMCD = "Refused"},
New T1 With {.CAMPAIGNID = 2, .TERMCD = "Accepted"},
New T1 With {.CAMPAIGNID = 3, .TERMCD = "Refused"},
New T1 With {.CAMPAIGNID = 3, .TERMCD = "Accepted"}
}
Dim tbl2 As New List(Of T2) From {
New T2 With {.CAMPAIGNID = 1, .CALLS = 2},
New T2 With {.CAMPAIGNID = 1, .CALLS = 4},
New T2 With {.CAMPAIGNID = 2, .CALLS = 8},
New T2 With {.CAMPAIGNID = 2, .CALLS = 16},
New T2 With {.CAMPAIGNID = 3, .CALLS = 32},
New T2 With {.CAMPAIGNID = 3, .CALLS = 64}
}
Dim query =
From t1 In tbl1
Join t2 In tbl2 On t1.CAMPAIGNID Equals t2.CAMPAIGNID
Group By t1.CAMPAIGNID Into Group
Select New With {
.id = CAMPAIGNID,
.CALLS = Group.Sum(Function(a) a.t2.CALLS),
.count = Group.Count(Function(a) a.t1.TERMCD = "Refused")
}
For Each q In query
Console.WriteLine($"ID:{q.id}, CALLS:{q.CALLS}, COUNT:{q.count}")
Next
ID:1, CALLS:12, COUNT:2
ID:2, CALLS:48, COUNT:2
ID:3, CALLS:192, COUNT:2
In the case of ID 3, the CALLS sum of 192 is correct because the sum has no constraint of TERMCD = "Refused", so it counts all. And the count of 2 is accurate because has a constraint where TERMCD = "Refused".
Just looking at the group, without aggregation, let's see what is returned
Dim query =
From t1 In tbl1
Join t2 In tbl2 On t1.CAMPAIGNID Equals t2.CAMPAIGNID
Group By t1.CAMPAIGNID Into Group
For Each q In query
Console.WriteLine($"ID:{q.CAMPAIGNID}")
For Each g In q.Group
Console.WriteLine($"t1.CAMPAIGNID:{g.t1.CAMPAIGNID}, t1.TERMCD:{g.t1.TERMCD}, t2.CAMPAIGNID:{g.t2.CAMPAIGNID}, t2.CALLS:{g.t2.CALLS}")
Next
Next
ID:1
t1.CAMPAIGNID:1, t1.TERMCD:Refused, t2.CAMPAIGNID:1, t2.CALLS:2
t1.CAMPAIGNID:1, t1.TERMCD:Refused, t2.CAMPAIGNID:1, t2.CALLS:4
t1.CAMPAIGNID:1, t1.TERMCD:Accepted, t2.CAMPAIGNID:1, t2.CALLS:2
t1.CAMPAIGNID:1, t1.TERMCD:Accepted, t2.CAMPAIGNID:1, t2.CALLS:4
ID:2
t1.CAMPAIGNID:2, t1.TERMCD:Refused, t2.CAMPAIGNID:2, t2.CALLS:8
t1.CAMPAIGNID:2, t1.TERMCD:Refused, t2.CAMPAIGNID:2, t2.CALLS:16
t1.CAMPAIGNID:2, t1.TERMCD:Accepted, t2.CAMPAIGNID:2, t2.CALLS:8
t1.CAMPAIGNID:2, t1.TERMCD:Accepted, t2.CAMPAIGNID:2, t2.CALLS:16
ID:3
t1.CAMPAIGNID:3, t1.TERMCD:Refused, t2.CAMPAIGNID:3, t2.CALLS:32
t1.CAMPAIGNID:3, t1.TERMCD:Refused, t2.CAMPAIGNID:3, t2.CALLS:64
t1.CAMPAIGNID:3, t1.TERMCD:Accepted, t2.CAMPAIGNID:3, t2.CALLS:32
t1.CAMPAIGNID:3, t1.TERMCD:Accepted, t2.CAMPAIGNID:3, t2.CALLS:64
Now it should be clear why you see what you see.
Sum with no constraint:
t1.CAMPAIGNID:3, t1.TERMCD:Refused, t2.CAMPAIGNID:3, t2.CALLS:32
t1.CAMPAIGNID:3, t1.TERMCD:Refused, t2.CAMPAIGNID:3, t2.CALLS:64
t1.CAMPAIGNID:3, t1.TERMCD:Accepted, t2.CAMPAIGNID:3, t2.CALLS:32
t1.CAMPAIGNID:3, t1.TERMCD:Accepted, t2.CAMPAIGNID:3, t2.CALLS:64
Count where a.t1.TERMCD = "Refused"
t1.CAMPAIGNID:3, t1.TERMCD:Refused, t2.CAMPAIGNID:3, t2.CALLS:32
t1.CAMPAIGNID:3, t1.TERMCD:Refused, t2.CAMPAIGNID:3, t2.CALLS:64
I'm not sure what you want to see, but this should at least show you that your output is accurate.
Now that we got that out of the way, your SQL query is below
SELECT DISTINCT A.CAMPAIGNID
, RFJOIN.RF
, CCJOIN.CC
, CALLSJOIN.CALLS
, CALLSJOIN.CALLS / RFJOIN.RF as Calculated
FROM tbl1 A
inner join (SELECT CAMPAIGNID
, count() as RF
FROM tbl1
WHERE TERMCD='Refused'
GROUP by CAMPAIGNID) RFJOIN on RFJOIN.CAMPAIGNID = a.CAMPAIGNID
inner join (SELECT CAMPAIGNID
, count() as CC
FROM tbl1
WHERE TERMCD LIKE '%Con%'
GROUP by CAMPAIGNID) CCJOIN on CCJOIN.CAMPAIGNID = a.CAMPAIGNID
inner join (SELECT CAMPAIGNID
, SUM(CALLS) as CALLS
FROM tbl2
GROUP by CAMPAIGNID) CALLSJOIN on CALLSJOIN.CAMPAIGNID = a.CAMPAIGNID
I would break it into the subqueries then join them together. You can almost copy the SQL syntax this way
Dim RFJ = From t1 In tbl1
Where t1.TERMCD = "Refused"
Group By t1.CAMPAIGNID Into Group
Select New With {CAMPAIGNID, .RF = Group.Count()}
Dim CCJ = From t1 In tbl1
Where t1.TERMCD.Contains("Con")
Group By t1.CAMPAIGNID Into Group
Select New With {CAMPAIGNID, .CC = Group.Count()}
Dim CAJ = From t2 In tbl2
Group By t2.CAMPAIGNID Into Group
Select New With {CAMPAIGNID, .CALLS = Group.Count()}
Dim query = From A In tbl1
Join RFJOIN In RFJ On RFJOIN.CAMPAIGNID Equals A.CAMPAIGNID
Join CCJOIN In CCJ On CCJOIN.CAMPAIGNID Equals A.CAMPAIGNID
Join CALLSJOIN In CAJ On CALLSJOIN.CAMPAIGNID Equals A.CAMPAIGNID
Select New With {A.CAMPAIGNID, RFJOIN.RF, CCJOIN.CC, CALLSJOIN.CALLS, .Calculated = CALLSJOIN.CALLS / RFJOIN.RF}
Dim queryDistinct = query.Distinct()
' or in one big ugly query :)
Dim fullQuery = (From A In tbl1
Join RFJOIN In From t1 In tbl1
Where t1.TERMCD = "Refused"
Group By t1.CAMPAIGNID Into Group
Select New With {CAMPAIGNID, .RF = Group.Count()} On RFJOIN.CAMPAIGNID Equals A.CAMPAIGNID
Join CCJOIN In From t1 In tbl1
Where t1.TERMCD.Contains("Con")
Group By t1.CAMPAIGNID Into Group
Select New With {CAMPAIGNID, .CC = Group.Count()} On CCJOIN.CAMPAIGNID Equals A.CAMPAIGNID
Join CALLSJOIN In From t2 In tbl2
Group By t2.CAMPAIGNID Into Group
Select New With {CAMPAIGNID, .CALLS = Group.Count()} On CALLSJOIN.CAMPAIGNID Equals A.CAMPAIGNID
Select New With {A.CAMPAIGNID, RFJOIN.RF, CCJOIN.CC, CALLSJOIN.CALLS, .Calculated = CALLSJOIN.CALLS / RFJOIN.RF}).Distinct()

Answer 2

It truly is a case of me making it WAY more complicated then it was and not knowing enough to ask the right questions. I wanted to mimic the subquery concept in TSQL but I didn't explain it well. What threw me was the LINQ syntax which seems to always be my downfall with LINQ. I know what I wanted to do I just didn't know how to do it the LINQ way. Thanks to you all who have helped especially #djv.
Dim query =
From t1 In tbl1 Group t1 By t1.CAMPAIGNID Into grp = Group
Join t2 In (From p In tbl2 Group p By p.CAMPAIGNID Into g = Group
Select New With {
.CAMPAIGNID = CAMPAIGNID,
.CALLS = CType(g.Sum(Function(d) d.CALLS), Integer?).GetValueOrDefault(0),
})
On CAMPAIGNID Equals t2.CAMPAIGNID
Select New With {
.CampaignID = CAMPAIGNID,
.RF = CType(grp.Count(Function(b) b.TERMCD = "Refused"), Integer?).GetValueOrDefault(0),
.CALLS = t2.CALLS
}