Take the 2-minute tour ×
Database Administrators Stack Exchange is a question and answer site for database professionals who wish to improve their database skills and learn from others in the community. It's 100% free, no registration required.

This question is similar to Running total with count?, but please allow me to explain some further twists on the issue.

I'm using SQL Server 2008, so the cursor option described by Aaron Bertrand seems to be the most promising in terms of speed.

What's different here, though, is I have to take into account two dates for a single item. So, given an OrderID item within a table of Orders, the OrderID has an Opened Date and a Closed Date. The Opened Date is always populated, but the Closed Date could be NULL.

OrderID  OpenedDate    ClosedDate
654554    12/1/2011     5/4/2012
678451    12/4/2011     3/2/2012
679565    12/8/2011     5/21/2012
701541    5/23/2012     NULL

I need to - efficiently - get back how many Orders had what we could term an "Open" status on any given date within a date range. The date range could span a couple of years. (Yes, I do have a reference table of sequential dates.)

Date         CountOfOpenOrders
12/1/2011    175
12/2/2011    178
12/3/2011    195
12/4/2011    192
12/5/2011    191
share|improve this question
What does a running total means in business terms in this case? I can see a business meaning for the number of open accounts at any given date but not for a running total. –  Emmad Kareem Sep 5 '12 at 0:50
@EmmadKareem: The OP may be referring to a running total question simply because they could see a connection between the two problems, but I think this one isn't fundamentally a running total problem. (Although it can be turned into one, as AlexKuznetsov has demonstrated.) –  Andriy M Sep 5 '12 at 13:45
add comment

3 Answers

If your priority is speed of selects, the following approach allows for extremely fast selects. Instead of storing a period, you want to store two events (period start and period end). Change column is 1 when the period begins, and is -1 when the period ends. If more than one event occurs on the same day, they must have different EventNumberPerDay. RunningTotal is the number of open peroids after the event has happened:

CREATE TABLE dbo.Events(
 ChangeDate DATE NOT NULL,
 EventNumberPerDay INT NOT NULL,
 RunningTotal INT NOT NULL);

INSERT dbo.Events
        ( PeriodId ,
          Change ,
          ChangeDate ,
-- first period begins    
VALUES  ( 1 , 1, '20120801', 1, 1),
-- second period begins on the same day
(2, 1, '20120801', 2, 2),
-- third period begins
(3,1,'20120803',1, 3),
-- second period ends on the same day
(2,-1,'20120803',2, 2),
-- fourth period begins
-- fourth period ends
-- first period ends
(1, -1, '20120808',1,  1),
-- third period ends
(3, -1, '20120809',1,  0);

Also you need a calendar table:

CREATE TABLE dbo.Calendar([Date] DATE NOT NULL);

INSERT INTO dbo.Calendar([Date])

Once that is accomplished, then your select is very simple and very fast:

SELECT  [Date] ,
        coalesce(RunningTotal, 0) AS NumOpenIntervals
FROM    dbo.Calendar
        OUTER APPLY ( SELECT TOP ( 1 )
                      FROM      dbo.Events
                      WHERE     ChangeDate <= [Date]
                      ORDER BY  ChangeDate DESC, EventNumberPerDay DESC
                    ) AS t

Of course, this query is only correct if the data in Events table is valid. We can use constraints to ensure 100% data integrity. I can explain how if you are interested.

Another alternative is to just load your raw data, your periods, into a client application - your problem is absolutely trivial in C++/C#/Java.

Yet another approach is to use an RDBMS with fast cursors such as Oracle - that will allow you to just write a simple cursor and enjoy good performance, but still not always as good as my first solution.

share|improve this answer
add comment

I don't think you question is really about running totals. In this case you could just use an outer join between your "reference table of sequential dates" (called Calendar in the query below) and the Orders table, group the result set by Calendar.Date and count matching Orders rows:

  COUNT(o.OrderID) AS CountOfOpenOrders  /* or COUNT(o.AnyOtherColumn) */
FROM Calendar c
LEFT JOIN Orders o ON c.Date BETWEEN o.OpenedDate AND o.ClosedDate
-- or, perhaps, ON c.Date >= o.OpenedDate
--             AND c.Date <  o.ClosedDate
share|improve this answer
The JOIN you suggest does return the "CountOfOpenOrders" I'm looking for, but the more substantial the number of Orders becomes, the more the query slows down. –  mg1075 Sep 5 '12 at 15:09
add comment

Another way to ensure very fast selects is to store periods as sequences of days - then we can select number of open intervals directly from an indexed view dbo.ConcurrentPeriodsByDay, which is as fast as it goes.

CREATE TABLE dbo.PeriodsByDay(PeriodId INT NOT NULL,

INSERT INTO dbo.PeriodsByDay(PeriodId, AsOfDate)
VALUES(1, '20120801'),
(1, '20120802'),
(1, '20120803'),
(2, '20120802'),
(2, '20120803');

CREATE VIEW dbo.ConcurrentPeriodsByDay
SELECT AsOfDate, count_big(*) AS ConcurrentPeriodsByDay
FROM dbo.PeriodsByDay

    ON dbo.ConcurrentPeriodsByDay(AsOfDate);

SELECT * FROM dbo.ConcurrentPeriodsByDay WITH(NOEXPAND);

As in my previous solution, we can use constraint to ensure the integrity of data (the sequence of dates must have no gaps or duplicates). It is relatively easy.

share|improve this answer
add comment

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.