solving sql server problems for millions of dbas and developers since 2006

SQL Product Highlight

Idera - SQL diagnostic manager

Identify and resolve SQL Server problems before they happen

Monitor and manage SQL Servers enterprise-wide
Find and fix performance bottlenecks
Analyze performance over time

Learn more!

Using Derived Tables to Simplify the SQL Server Query Process

By: Greg Robidoux | Read Comments (8) | Related Tips: More > Query Optimization

Free SQL Server Learning - Using SQL Server DMVs to Help Improve Performance

Problem
Sometimes querying data is not that simple and there may be the need to create temporary tables or views to predefine how the data should look prior to its final output. Unfortunately there are problems with both of these approaches if you are trying to query data on the fly.

With the temporary tables approach you need to have multiple steps in your process, first to create the temporary table, then to populate the temporary table, then to select data from the temporary table and lastly cleanup of the temporary table.

With the view approach you need to predefine how this data will look, create the view and then use the view in your query. Granted if this is something that you would be doing over and over again this might make sense to just create a view, but let's look at a totally different approach.

Solution
With SQL Server you have the ability to create derived tables on the fly and then use these derived tables within your query. In concept this is similar to creating a temporary table and then using the temporary table in your query, but the approach is much simpler, because it can all be done in one step.

Let's take a look at an example where we query the Northwind database to try to find out how many customers fall into various categories based on sales. The categories that we have predefined are as follows:

Total Sales between 0 and 5,000 = Micro
Total Sales between 5,001 and 10,000 = Small
Total Sales between 10,001 and 15,000 = Medium
Total Sales between 15,001 and 20,000 = Large
Total Sales > 20,000 = Very Large

There are several ways that this data can be pulled, but let's look at an approach using a derived table.

The first step is to find out the total sales by each customer, which can be done with the following statement.

SELECT   o.CustomerID, 
         SUM(UnitPrice * Quantity) AS TotalSales 
FROM     [Order Details] AS od 
         INNER JOIN Orders AS o 
           ON od.OrderID = o.OrderID 
GROUP BY o.CustomerID

This is a partial list of the output:

CustomerID	TotalSales
ALFKI	4596.2000
ANATR	1402.9500
ANTON	7515.3500
...
WOLZA	3531.9500

The next step is to classify the TotalSales value into the OrderGroups that were specified above:

SELECT   o.CustomerID, 
         SUM(UnitPrice * Quantity) AS TotalSales, 
         CASE  
           WHEN SUM(UnitPrice * Quantity)  
               BETWEEN 0 AND 5000 THEN 'Micro' 
           WHEN SUM(UnitPrice * Quantity)  
               BETWEEN 5001 AND 10000 THEN 'Small' 
           WHEN SUM(UnitPrice * Quantity)  
               BETWEEN 10001 AND 15000 THEN 'Medium' 
           WHEN SUM(UnitPrice * Quantity)  
               BETWEEN 15001 AND 20000 THEN 'Large' 
           WHEN SUM(UnitPrice * Quantity)  
               > 20000 THEN 'Very Large' 
         END AS OrderGroup 
FROM     [Order Details] AS od 
         INNER JOIN Orders AS o  
          ON od.OrderID = o.OrderID 
GROUP BY o.CustomerID

This is a partial list of the output:

CustomerID	TotalSales	OrderGroup
ALFKI	4596.2000	Micro
ANATR	1402.9500	Micro
ANTON	7515.3500	Small
...
WOLZA	3531.9500	Micro

The next step is to figure out how many customers fit into each of these groups and this is where the derived table comes into play. Take a look at the following query which uses a derived table called OG. What we are doing here is using the same query from the step above, but calling this derived table OG. Then we are selecting data from this derived table for our final output just like we would with any other query. All of the columns that are created in the derived table are now available for our final query.

SELECT   OG.OrderGroup, 
         COUNT(OG.OrderGroup) AS OrderGroupCount 
FROM     (SELECT   o.CustomerID, 
                   SUM(UnitPrice * Quantity) AS TotalSales, 
                   CASE  
                     WHEN SUM(UnitPrice * Quantity)  
                       BETWEEN 0 AND 5000 THEN 'Micro' 
                     WHEN SUM(UnitPrice * Quantity)  
                       BETWEEN 5001 AND 10000 THEN 'Small' 
                     WHEN SUM(UnitPrice * Quantity)  
                       BETWEEN 10001 AND 15000 THEN 'Medium' 
                     WHEN SUM(UnitPrice * Quantity)  
                       BETWEEN 15001 AND 20000 THEN 'Large' 
                     WHEN SUM(UnitPrice * Quantity)  
                       > 20000 THEN 'Very Large' 
                   END AS OrderGroup 
          FROM     [Order Details] AS od 
                   INNER JOIN Orders AS o 
                     ON od.OrderID = o.OrderID 
          GROUP BY o.CustomerID) AS OG 
GROUP BY OG.OrderGroup

This is the complete list of the output from the above query.

OrderGroup	OrderGroupCount
Large	10
Medium	11
Micro	33
Small	15
Very Large	20

Next Steps

Next time you run into a challenge of whether to create a temporary table or a view to produce the desired query take a look at using a derived table instead
Experiment with using derived tables, views and temporary tables to see what yields better performance results. There are several articles on the internet that have shown that using a derived table is faster then temporary tables, but try it for yourself.
Take a look at Common Table Expressions in SQL Server 2005

Last Update: 8/16/2006

About the author

Become a paid author

Comments and Feedback:

Tuesday, December 21, 2010 - 12:33:41 PM - Jackie Bird	Read The Tip
Tried your derived table example - but when you have multiple fields to pull you have to include all of them in the group by clause - moreover the AS for the inner group by doesn't work - it throws an error and when i remove it the sql executes

Tuesday, December 21, 2010 - 1:31:54 PM - Greg Robidoux

Read The Tip

Hi Jackie, what version of SQL Server are you using?

Since all of the data is GROUPed you will need to include any column that is not a summary such as AVG, SUM, COUNT, etc... in the GROUP BY clause.

Also, I am not sure why the AS for the derived table did not work. This is not really necessary. I tried this both with and without the AS and it worked fine both ways.

Tuesday, June 19, 2012 - 7:37:13 AM - Sulthanul Arifeen	Read The Tip
Derived tables only helpful for small amount data tranction. For huge data transaction with calculation, temp table is better option. My suggestion is to use temporary table whenever possible only keeping the locking issue under consideration.

Sunday, March 03, 2013 - 11:52:24 AM - Frank

Read The Tip

Im am trying to create a sql query that will return the following (1) from the table snippet (2) I am using this (3):

(1)

EmdDescription

10D110 - CHEST PAIN - D1 Severe Respiratory distress

10D210 - CHEST PAIN - D2 Not alert

10D310 - CHEST PAIN - D3 Clammy

(2)

EmdDescription

10D110 - CHEST PAIN - D1 Severe Respiratory distress

10D210 - CHEST PAIN - D2 Not alert

10D310 - CHEST PAIN - D3 Clammy

11A111 - CHOKING - A1 Not choking now

11D111 - CHOKING - D1 Not alert

11D211 - CHOKING - D2 Abnormal breathing

11011 - CHOKING - E1 Choking verified or ineffective breathing

12A112 - CONVULSIONS SEIZURES - A1 Not seizing now regularly

12B112 - CONVULSIONS SEIZURES - B1 Breathing regular not verified < 35 yrs

SELECT *

FROM EmdCodes

WHERE Description ="%CHESTPAIN%";

Any help is appreaciated.

Wednesday, March 06, 2013 - 2:50:38 PM - George	Read The Tip
Any thoughts on whether using a Common Table Expression is better than a derived table?

Monday, April 01, 2013 - 11:01:43 AM - RHG	Read The Tip
Very helpful, thanks a lot

Monday, April 01, 2013 - 3:35:27 PM - Greg Robidoux

Read The Tip

@Frank - I think you just need a space in your query:

SELECT *

FROM EmdCodes

WHERE Description ="%CHEST PAIN%";

Monday, April 01, 2013 - 3:36:42 PM - Greg Robidoux	Read The Tip
@George - not totally sure if a CTE will perform better. The best option is to test both methods and see what is eaiser to setup and also what performs better.

Sponsor Information

"Amazing, Amazing, Amazing! SQL doctor is truly one of the most powerful tools I have seen."

SQL Monitor: prioritize your SQL Server workload with easy-to-use performance monitoring

Is “blocking” a bad word at your company? Contact the Edgewood SQL Server Consultants for the resolution.

Secure column & whole database on all versions and editions of SQL Server with NetLib’s TDE

Optimizing SQL Server performance can be a daunting task. Or is it?