Siddhartha Gautama, the Buddha, taught us to understand that the key to enlightenment is following the Middle Path. And today I learned a valuable lesson in extremes. You can file this one in the “Doh! Wrong again!” category…
A fairly common question on SQL Server forums is, “how can I get the running sum of the data in this column?” Being the fan of set-based queries that I am, I always answer the exact same way. I show the person asking the question how to do a self-join on the grouped column, getting all of the “previous” values to create a running sum. The following example shows how you might do this against the AdventureWorks Production.TransactionHistory table:
[sql]SELECT
TH1.TransactionID,
TH1.ActualCost,
SUM(TH2.ActualCost) AS RunningTotal
FROM Production.TransactionHistory TH1
JOIN Production.TransactionHistory TH2 ON TH2.TransactionID <= TH1.TransactionID
GROUP BY TH1.TransactionID, TH1.ActualCost
ORDER BY TH1.TransactionID
[/sql] Pretty simple query. For each row of the “TH1” alias, every row with a lesser-or-equal TransactionID will be summed. Thereby creating a running total for every row of the table. Note, I’ve used the IDENTITY column instead of the date column. I’d generally suggest not doing so because, e.g., you might need to insert some post-dated rows at some point and relying on the IDENTITY for a time sequence will thereby not work. But in this case it’s a lazy solution because the TransactionDate column isn’t indexed, and it’s also not unique. I want to test a lot of rows (TransactionHistory has around 113,000), but I don’t want to skew the test by forcing a table scan on every iteration!
But I digress. The point is, I’ve given this answer more than a few times and, well, I’d like to apologize. Just now I went ahead and ran this query on my powerful test server–err, my laptop.
As you might guess, since I’m performance-minded I also happen to be extremely impatient–so I went ahead and killed the query at the five-minute mark. SSMS’s result grid showed the first 26,568 rows, so obviously there was a long way to go to hit that 113,000 mark. And with an estimated cost of 38,086 for the query, I can’t say I’m surprised.
A few moments of head scratching and the following re-write was issued:
[sql]SELECT
TH1.TransactionID,
TH1.ActualCost,
(
SELECT SUM(TH2.ActualCost)
FROM Production.TransactionHistory TH2
WHERE TH2.TransactionID <= TH1.TransactionID
) AS RunningTotal
FROM Production.TransactionHistory TH1
ORDER BY TH1.TransactionID
[/sql] With an estimated cost of only 6,630, I had high hopes for this one. Alas, once again I was forced to cancel the query at the five-minute mark. 27,683 rows. Not much better, I’m afraid. And, as an aside, I’m starting to wonder about these estimated costs. But that’s another post for another day.
So where am I going with all of this? Well, there’s a reason I haven’t given any indication up until this point in the post. You see, it’s utterly painful to write this, but…
In this case, a cursor is faster.
Yes, I said it. That evil construct which we as database developers despise, the cursor. Thanks to Paul Nielsen, who revealed this ugly fact to me in a conversation today, I was forced to test this for myself (hoping to prove him wrong, of course). Which is why I started playing around with the solution that I’ve given so many times on forums. Unfortunately, he is correct.
My next test query, using the first cursor I’ve written in several years:
[sql]DECLARE RunningTotalCursor
CURSOR LOCAL FAST_FORWARD FOR
SELECT TransactionID, ActualCost
FROM Production.TransactionHistory
ORDER BY TransactionID
OPEN RunningTotalCursor
DECLARE @TransactionID INT
DECLARE @ActualCost MONEY
DECLARE @RunningTotal MONEY
SET @RunningTotal = 0
DECLARE @Results TABLE
(
TransactionID INT NOT NULL PRIMARY KEY,
ActualCost MONEY,
RunningTotal MONEY
)
FETCH NEXT FROM RunningTotalCursor
INTO @TransactionID, @ActualCost
WHILE @@FETCH_STATUS = 0
BEGIN
SET @RunningTotal = @RunningTotal + @ActualCost
INSERT @Results
VALUES (@TransactionID, @ActualCost, @RunningTotal)
FETCH NEXT FROM RunningTotalCursor
INTO @TransactionID, @ActualCost
END
CLOSE RunningTotalCursor
DEALLOCATE RunningTotalCursor
SELECT *
FROM @Results
ORDER BY TransactionID
[/sql] What’s really unfortunate about the cursor approach is that you need to use a temporary table if you want to return a single result set to the client. I figured the additional I/O due to the temp table would balance any improvement gains from the cursor approach, thereby rendering my forum responses correct, and Paul wrong. Well, 14 seconds and 113,443 rows later, SSMS and my laptop declared Paul the undisputed Champion of the Cursor.
This cursor makes a lot of sense in this case. The set-based query works by looping over each row of the table, taking a sum of every previous row. So for the 10th row, 10 previous rows also need to be visited. For the 1000th row, 1000 previous rows need to be visited. And so on. The larger the set gets, the worse performance will be–and that’s not going to be a merely linear decrease in performance. Think about this: Using the set-based method to find the running sum over a set of 100 rows, 5050 total rows need to be visited. For a set of 200 rows, the query processor needs to visit 20100 total rows — a four-fold increase in the amount of work that must be done to satisfy the query (O((N^2)/2), for those who are a bit more algorithmically minded.)
The cursor, on the other hand, needs to visit each row exactly once (O(N)). By maintaining the running count in a variable, there is no need to re-visit previous rows. And as my laptop was so happy to show me, the I/O cost due to the temp table does not overshadow the performance improvement of having to visit so many less rows.
So what have we learned today? In my set-based singlemindedness I failed to realize that the cursor does, indeed, have utility. Everything in moderation.
Next steps? I get the feeling that this can be made even faster by employing a CLR routine. Pull the data into a DataReader and loop over that instead, which will completely eliminate the need for a temporary table. Watch for that experiment coming to this space soon.
And next time you hear someone mention how horrible cursors are, remind that person that there is a time and place for everything (and it’s called college).
There is still overhead in the cursor that you can escape. On way to perform this running total style calculation in a "set" oriented way that avoid the heavy recursion of the first two examples is…
SET NOCOUNT ON
DECLARE @hist TABLE ( TransactionID INT PRIMARY KEY CLUSTERED, ActualCost money, RunningTotal money )
DECLARE @total money; SET @total = 0.0
DECLARE @start datetime, @stop datetime
SET @start = GETDATE()
INSERT INTO @hist ( TransactionID, ActualCost, RunningTotal )
SELECT TransactionID, ActualCost, 0.0
FROM Production.TransactionHistory
ORDER BY TransactionID
UPDATE @hist
SET @total = RunningTotal = ActualCost + @total
SET @stop = GETDATE()
PRINT ‘Time to calculate running total is ‘ + CONVERT(VARCHAR(10), DATEDIFF(ms, @start, @stop)) + ‘ms’
SELECT TransactionID, ActualCost, RunningTotal
FROM @hist
Sorry about the double post here, but I found it very interesting to turn on IO statistics for both so you can see the main differences. I took some liberties in moving around declarations in the original cursor code to better accommodate timing the calculation of the running total.
If you run these, you will immediately notice the difference…
/***** ORIGINAL *****/
SET NOCOUNT ON
SET STATISTICS IO ON
DECLARE RunningTotalCursor
CURSOR LOCAL FAST_FORWARD FOR
SELECT TransactionID, ActualCost
FROM Production.TransactionHistory
ORDER BY TransactionID
DECLARE @TransactionID INT
DECLARE @ActualCost MONEY
DECLARE @RunningTotal MONEY
SET @RunningTotal = 0
DECLARE @Results TABLE
(
TransactionID INT NOT NULL PRIMARY KEY,
ActualCost MONEY,
RunningTotal MONEY
)
DECLARE @start datetime, @stop datetime
SET @start = getdate()
OPEN RunningTotalCursor
FETCH NEXT FROM RunningTotalCursor
INTO @TransactionID, @ActualCost
WHILE @@FETCH_STATUS = 0
BEGIN
SET @RunningTotal = @RunningTotal + @ActualCost
INSERT @Results
VALUES (@TransactionID, @ActualCost, @RunningTotal)
FETCH NEXT FROM RunningTotalCursor
INTO @TransactionID, @ActualCost
END
CLOSE RunningTotalCursor
DEALLOCATE RunningTotalCursor
SET @stop = getdate()
PRINT ‘Time to calculate running total is ‘ + convert(varchar(10), datediff(ms, @start, @stop)) + ‘ms’
SELECT *
FROM @Results
ORDER BY TransactionID
/***** ALTERNATE *****/
SET NOCOUNT ON
SET STATISTICS IO ON
DECLARE @hist table ( TransactionID int primary key clustered, ActualCost money, RunningTotal money )
DECLARE @total money; SET @total = 0.0
DECLARE @start datetime, @stop datetime
SET @start = getdate()
INSERT INTO @hist ( TransactionID, ActualCost, RunningTotal )
SELECT TransactionID, ActualCost, 0.0
FROM Production.TransactionHistory
ORDER BY TransactionID
UPDATE @hist
SET @total = RunningTotal = ActualCost + @total
SET @stop = getdate()
PRINT ‘Time to calculate running total is ‘ + convert(varchar(10), datediff(ms, @start, @stop)) + ‘ms’
SELECT TransactionID, ActualCost, RunningTotal
FROM @hist
I have tried to come up with a solution that didn’t require a cursor to calculate a running total,the following solution ran in 57 secs on my machine and does not use a cursor.
I have used the production.TransactionHistory table in the AdventureWorks database.
Declare @iCounter int
Declare @i int
Declare @mintrans int, @TransId int, @Count1 int
SET NOCOUNT ON;
Set @i = 1
Select @iCounter = Count(*) From production.TransactionHistory
Select @mintrans = Min(transactionId) From production.TransactionHistory
While @i <= @iCounter
Begin
Select @transId = TransactionId From production.TransactionHistory where Transactionid = @mintrans
Update th1
Set th1.RT = ISNULL(th1.ActualCost + th2.RT, th1.ActualCost)
From production.TransactionHistory th1
Join production.TransactionHistory th2 on th2.TransactionId = th1.TransactionId – 1
Where th1.TransactionId = @transId
Set @mintrans = @mintrans + 1
Set @i = @i+1
End
Select th1.TransactionId, th1.ActualCost, ISNULL(th1.ActualCost + th2.RT, th1.ActualCost) AS RT
From production.TransactionHistory th1
Join production.TransactionHistory th2 on th2.TransactionId = th1.TransactionId – 1
There are improvements that can be made to this – but this was my first attempt at trying to do this without a cursor, even though a look is used to perform the update statement, which is the first stage of this task.
Forgot to mention, that you will have to add a new column to the table, i have called it RT and it’s type is Money.
Hi Adam,
What about doing the Running Sum calculations in reporting application? Is it better idea?
I used SSRS with follwoing query
SELECT SalesOrderID,SalesOrderDetailID,
OrderQty
FROM AdventureWorks.Sales.SalesOrderDetail
and the RunningValue function – =RunningValue(Fields!OrderQty.Value,Sum,Nothing)
Just looking at it I see RunningValue function performing better than if I write T-SQL code.
What’s your thought?
Shivani: Absolutely! The application is 100%, without a doubt, the right place to do this kind of work. But sometimes we need to make exceptions even to 100% without a doubt rules 🙂
I know this is a really old article but if you ever want examples of where window functions in 2012 really clean up it’s this sort of thing.
SELECT
TH1.TransactionID,
TH1.ActualCost,
TotalAmount = SUM(TH1.ActualCost) OVER (ORDER BY TH1.TransactionID ROWS UNBOUNDED PRECEDING)
FROM Production.TransactionHistory TH1
ORDER BY TH1.TransactionID
Runs in under 1 second.
One way of solving this without using cursors is to use a recursive Common Table Expression.
E.g. http://blog.sqltechie.com/2009/04/how-to-do-running-sum-using-cte-sql.html
@Philip
I think Mike’s comment is where the discussion ends, in the post-SQL Server 2012 era. Window function enhancements have revolutionized the way we think about these kinds of problems.
Here’s a talk I did on the subject a while back: https://www.youtube.com/watch?v=YK-ufNpMeLU