SQL - Group By Into Separate Columns

Question

Using SQL Server 2008 R2. I'm not sure if this is possible, but I would like to group by two columns and take their differences and place them into new columns. It's probably easier to explain with an example:

BEFORE::

[DATE]      [ID]                                  [AMT] [TYPE]
2013-01-11  36374DCD-47FE-48D8-8E70-8D3B37385311    2   20
2013-01-11  36374DCD-47FE-48D8-8E70-8D3B37385311    10  21
2013-01-11  4434E2D5-1D08-45FA-AADF-F653BF9A0D97    4   20
2013-01-11  4434E2D5-1D08-45FA-AADF-F653BF9A0D97    4   21

AFTER::

[DATE]      [ID]                                 [AMT 20] [AMT 21]
2013-01-11  36374DCD-47FE-48D8-8E70-8D3B37385311    2       10
2013-01-11  4434E2D5-1D08-45FA-AADF-F653BF9A0D97    4       4

This is the SQL I have so far:

SELECT 
    CAST(TransDate AS DATE) AS [TransDate],ItemID,COUNT(TransactionTypeID) AS [TransAmt], TransactionTypeID
FROM 
    Transactions
WHERE 
    TransDate BETWEEN '2013-01-01 10:00:00' AND '2013-02-01 10:00:00'
    AND TransactionTypeID IN (20,21)
GROUP BY
    CAST(TransDate AS DATE),ItemID,TransactionTypeID

Any help is appreciated, thanks!

Answer 1

You can use an aggregate function with a CASE to covert the rows into columns:

SELECT CAST(TransDate AS DATE) AS [TransDate],
  ItemID,
  count(case when TransactionTypeID=20 then TransactionTypeID end) Amt_20,
  count(case when TransactionTypeID=21 then TransactionTypeID end) Amt_21
FROM Transactions
WHERE 
    TransDate BETWEEN '2013-01-01 10:00:00' AND '2013-02-01 10:00:00'
    AND TransactionTypeID IN (20,21)
GROUP BY CAST(TransDate AS DATE),ItemID;

Since you are using SQL Server this can also be written using the PIVOT function:

select TransDate,
  ItemId,
  [20] as Amt_20,
  [21] as Amt_21
FROM
(
  SELECT CAST(TransDate AS DATE) AS [TransDate],
    ItemID,
    TransactionTypeID
  FROM Transactions
  WHERE TransDate BETWEEN '2013-01-01 10:00:00' AND '2013-02-01 10:00:00'
    AND TransactionTypeID IN (20,21)
) d
pivot
(
  count(TransactionTypeID)
  for TransactionTypeID in ([20], [21])
) piv

Answer 2

bluefeet answer might be the best way to do it, but if your query gets more complex, you can always use as many UNIONS as you need :)

    SELECT 
        CAST(TransDate AS DATE) AS [TransDate],ItemID,COUNT(TransactionTypeID) AS [TransAmt], TransactionTypeID
    FROM 
        Transactions
    WHERE 
        TransDate BETWEEN '2013-01-01 10:00:00' AND '2013-02-01 10:00:00'
        AND TransactionTypeID IN (20)
    GROUP BY
        CAST(TransDate AS DATE),ItemID,TransactionTypeID
UNION
SELECT 
    CAST(TransDate AS DATE) AS [TransDate],ItemID,COUNT(TransactionTypeID) AS [TransAmt], TransactionTypeID
FROM 
    Transactions
WHERE 
    TransDate BETWEEN '2013-01-01 10:00:00' AND '2013-02-01 10:00:00'
    AND TransactionTypeID IN (21)
GROUP BY
    CAST(TransDate AS DATE),ItemID,TransactionTypeID