SQL Server : Joining two pivot tables - sql

Question

I'm trying to create join a table with itself. So for example below the table pivots based on the sum of D (Debits) and C (Credits) however I need to join the table with itself to add additional columns displaying count of an "D" entry and "C" plus two more additional columns showing the overall sum and overall count. How do I join the table below to create the additional columns?
Input table
GL_BU GL_Source GL_JE_Type GL_Amount Amount_Prefix
------------------------------------------------------------------
202 Payables Purchase Invoices 1234 C
202 Payables Purchase Invoices 123 D
202 Inventory Inventory 123 C
202 Payables Purchase Invoices 1234 C
Output Table
GL_BU GL_Source GL_JE_Type Amount D Amount C Count D Count C Total Count Total Amount
------------------------------------------------------------------------------------------
202 Spreadsheet XXXXX 1234 123 1 1 2 1357
202 Manual XXXXX 1234 123 2 2 4 1357
202 Manual XXXXX 1234 123 1 1 2 1357
202 Inventory XXXXX 1234 123 4 4 8 1357
202 Sales Order XXXXXX 1234 123 1 1 2 1357
Current Code
SELECT *
FROM
(SELECT
[GL_Business_Unit]
,[GL_Source]
,[GL_JE_Type]
,([GL_Amount])
,[Amount_Prefix]
FROM [03_rdm].[table_2013]) as t
Pivot(SUM([GL_Amount])
FOR [Amount_Prefix] IN (D,C)) AS pvt1
Current code link in SQLFiddle http://sqlfiddle.com/#!3/92369/2

Answer 1

Your sample data doesn't match your desired result so I'm guessing that this is what you need. You could use a PIVOT to get the result, but it seems that this would be much easier to get this using an aggregate function and some conditional logic via a CASE expression:
select
GL_BU,
GL_Source,
GL_JE_Type,
sum(case when Amount_Prefix = 'D' then GL_Amount else 0 end) Amount_D,
sum(case when Amount_Prefix = 'C' then GL_Amount else 0 end) Amount_C,
sum(case when Amount_Prefix = 'D' then 1 else 0 end) Count_D,
sum(case when Amount_Prefix = 'C' then 1 else 0 end) Count_C,
count(*) TotalCount,
sum(GL_Amount) TotalAmount
from table_2013
group by GL_BU, GL_Source, GL_JE_Type;
See SQL Fiddle with Demo