Q
Problem solve Get help with specific problems with your technologies, process and projects.

# Averages over a span of years -- Part 2

In  Part 1 of this answer we examined averages produced by the AVG function on the following table:

```subject | year | enrolled ----------+---------+------------- subject1 | 1998 | 20 subject1 | 1999 | 23 subject1 | 2000 | 16 subject2 | 1999 | 10 subject2 | 2000 | 21 subject3 | 2000 | 9
```

The averages for each subject were the same whether we supplied the missing years or not. Here we'll explore why, and how to work with NULLs and aggregates.

When we supplied the missing years and found that averages were not affected, we demonstrated that aggregate functions exclude NULLs. The average was calculated using the number of non-NULL values in each subject group.

To get averages calculated over all years in the given span, where enrollment rows are missing, we must assume that the number enrolled was zero. This may not be a valid assumption in all applications. Remember, NULL is not equal to any value, and in particular, NULL is not equal to zero, so we have to do something specific to make it work that way.

The obvious solution is to check when the column is NULL, and use zero instead, which is completely in line with the assumption we are making and is the best way to solve the problem. We simply use the COALESCE function:

```select allyears.subject , avg(enrolled) as avgamt , avg(coalesce(enrolled,0)) as avgamtzero from ( select distinct subject , 1998+i as theyear from integers , subjects where i between 0 and 2 ) as allyears left outer join subjects on allyears.subject = subjects.subject and allyears.theyear = subjects.theyear group by allyears.subject
subject avgamt avgamtzero subject1 19.67 19.67 subject2 15.50 10.33 subject3 9.00 3.00
```

COALESCE is a standard SQL function. If your database does not support it, look for an equivalent function like ISNULL or NVL.

We made the correct calculation over the span of years by generating rows that were missing. Was this the easiest way? Consider the following query:

```select allyears.subject , sum(enrolled) as sumamt , count(enrolled) as countamt , count(*) as countrows , avg(enrolled) as avgamt , avg(coalesce(enrolled,0)) as avgamtzero from ( select distinct subject , 1998+i as theyear from integers , subjects where i between 0 and 2 ) as allyears left outer join subjects on allyears.subject = subjects.subject and allyears.theyear = subjects.theyear group by allyears.subject
subject sumamt countamt countrows avgamt avgamtzero subject1 59 3 3 19.67 19.67 subject2 31 2 3 15.50 10.33 subject3 9 1 3 9.00 3.00
```

Notice that COUNT(*) counts rows, whether any particular column had NULLs or not. COUNT(*) ignores NULLs. This is the only exception to the rule that aggregate functions exclude NULLs, because COUNT(*) does not even look at columns.

So the other way of getting the desired averages is:

``` sum(enrolled) / count(*) as avgamtzero
```

I can't say whether this method is as efficient as AVG(COALESCE(xxx)), but my guess is that they are approximately the same.

Finally, there is one other approach. Instead of generating rows that were missing with a cross join to the integers table, just do this:

```select subject , sum(enrolled) / 3 as avgamt from subjects group by subject
subject avgamt subject1 19.67 subject2 10.33 subject3 3.00
```

Is this reasonable? Can we just plug 3 into the calculation for the range of years desired? The derived table subquery was fashioned using inspection to determine the range of years. In general, it's not that easy. The generation of missing rows using the cross join would have used additional subqueries to determine the first and last years, so that inspection would not be necessary. If you can do it by inspection, okay, but it's nice to know how to attack the general problem, too.

This was last published in November 2002

## Content

Find more PRO+ content and other member only offers, here.

#### Have a question for an expert?

Get answers from a TechTarget expert on whatever's puzzling you.

You will be able to add details on the next page.

#### Start the conversation

Send me notifications when other members comment.

## SearchDataManagement

• ### Blockchain data disruption aborning, IDC analyst says

Blockchain data technology disruption may be in the offing. IDC's Stewart Bond says architecture at the core of controversial ...

• ### Potent NoSQL architecture engaged for building new applications

Behind hyperbolic terms like 'digital transformation' are innovative systems of engagement. DataStax CEO Billy Bosworth joins ...

• ### Finance data governance program gets new start, broader focus

Businesses constantly need to evolve their programs for governing data. Nationwide's finance data governance team shares how it ...

• ### Data science and analytics methods must balance innovation, value

In this episode of the Talking Data podcast, find out how data science teams at Cisco and LinkedIn navigate the process of doing ...

• ### Cognitive computing technologies still deliver mixed results

Cognitive computing tools have come a long way in the last couple of years, but the notion of true cognitive businesses, built ...

• ### Deloitte report reveals the power of unstructured data analytics

The analysis of unstructured data and other so-called dark data types can deliver significant business value, according to a new ...

## SearchSAP

• ### Future of Ariba Network on display at SAP Ariba Live

SAP Ariba Live shows enhancements to procurement software and looks at the future of the SAP Ariba Network, including machine ...

SAP Vora has been updated to include features that make it easier to deploy and use to get insights from Hadoop big data; SAP IBP...

• ### Chatbots provide faster self-service on SAP systems

The SAP SuccessFactors partnership with collaboration software vendor Slack is just one of many efforts to make little robotic ...

## SearchSQLServer

• ### Use these commands to deploy SQL Server Docker containers

Containerized apps are everywhere, but you can also run databases in containers. Follow these commands to customize SQL Server ...

• ### Redgate tools help bring database DevOps to Visual Studio 2017

So far, data has been left out of the DevOps discussion. But DevOps for databases could grow if tools like Redgate's bundle ...

• ### Conquer real-time operational analytics in SQL Server 2016

Analytics helps improve business operations, and SQL Server 2016 users can master it -- in real time, using operational data in a...

## TheServerSide

• ### Does reading XML while writing JSON make me a bad person?

I'm always writing JSON code, loving its flexibility and forgiveness. But when I call an external service, I want to be reading ...

• ### Stop adding web UI frameworks like JSR-371 to the Java EE spec

Java EE's Servlet and JSP API provide a fine foundation for handling a web-based request-response cycle. Adding UI frameworks ...

• ### Declarative Jenkins pipeline the latest new feature in DevOps tooling

Looking to get more out of your DevOps tooling? The declarative Jenkins pipeline helps take continuous integration and continuous...

## SearchDataCenter

• ### IoT certifications land on data center admins' to-do list

IoT is still an emerging field, but data center admins should at least consider certification options to make their resume stand ...

• ### Manage and optimize IBM z Systems software costs

Some mainframe users find it tough to navigate IBM's sub-capacity pricing model. Consider tools like SCRT to more efficiently ...

• ### Find the right data center cooling systems for hyper-converged

Hyper-converged infrastructure can cause new data center cooling challenges. Before implementation, determine which temperatures ...

## SearchContentManagement

• ### Seven features to consider when picking enterprise collaboration tools

As collaboration needs grow, more options are beginning to sprout up. Here are the features to think about when selecting the ...

• ### Microsoft Teams joins growing business chat software market

The general release of Microsoft's latest product puts Slack square in its scope, with hopes of taking a chunk from the business ...

Headless CMS can be a difficult pivot for dyed-in-the-wool legacy shops, but remixing content in this new model with RESTful APIs...

## SearchFinancialApplications

• ### Learning management platform helps keep Unisys employees up to date

The vast majority of employees at Unisys said they were more effective at their jobs after completing training on the cloud-based...

• ### Report rates e-sourcing, spend analysis and contract management tools

Gartner Magic Quadrant finds plenty of room for growth in market for cloud-based strategic sourcing application suites that can ...

• ### Benefits administration systems that use analytics liked by employers

HR tech systems are increasingly using analytics in benefits administration, including using claims data; meanwhile, APIs are ...

Close