I ran across a good article from Glenn Berry about reasons to upgrade to SQL Server 2017. In the piece, Glenn talks about the fact that SQL Server 2008 and 2008 R2 will fall out of Extended Support in 2019. That's about the time that SQL Server 2014 falls out of mainstream support, which is the support that most of us have. SQL Server 2012 has fallen out of mainstream support already. While I don't worry too much about support, some of you may, especially if you are in a regulated industry. Or you're in the EU, in which case the lack of support might be an issue if there is ever some security breach.

The upgrade question is one that most of us ask ourselves constantly. Many of us use plenty of different software packages, and regularly need to decide if we want to pay for a new version of Windows, Office, SQL, SAP, Dynamics, etc. In some cases we may find valid reasons that will give us a nice ROI and are worth the cost. In others, we may wonder if we really need some new feature? I certainly find that in some software, there isn't a compelling reason to upgrade to every new version.

These days so much software is being released at a rapid pace that we regularly have to make decisions. I get a new version of Visual Studio Code, Evernote and TweetDuck, etc. every few weeks. Often I skip these upgrades because of convenience, and only upgrade rarely to prevent being too far behind. That's a concern because if your software is too old, sometimes the upgrade is painful.

At Redgate, we release patches and enhancements every few weeks, though we don't expect many customers to necessarily upgrade more than once a quarter. There are monthly releases of SSMS and SQL Operations Studio, which I may or may not apply. Often I don't because they're an interruption to my work. Some of us rent software, like visualstudio.com, and we get upgrades whether we want them or not, though often we can use an older version for some period of time.

Many of these decisions aren't that impactful to our work as they are minor changes with little or no cost. Some aren't, and these are the ones that many technical professionals will make recommendations on whether to proceed or not. The database platform is certainly a significant decision as the costs often can significantly impact a budget. This can be especially true for mission critical applications that might require more than one server to meet HA requirements.

Do you tend to lean towards or away from upgrades? I've tended to lean away, with my default answer being I don't want to upgrade. That's if I don't have a compelling reason. If things are roughly equal, I'll stick with what works. That doesn't mean that I don't evaluate the new version and spend some time trying to determine what features or functions might be valuable. Certainly if any of Glenn's features are useful, you might consider moving to SQL Server 2017. If you're coming from SQL Server 2014 or earlier, than you ought to really look at the changes in SQL Server 2016, which are significant as well.

We recently moved SQLServerCentral to SQL Server 2017. We had been on an older OS, with limited .NET support, and when we made the decision, we decided to go ahead and choose the latest version. That's certainly a decision I do endorse. If I'm leaving 2012, 2008, or some earlier version, why stop at SQL Server 2014 or 2016. Go ahead and get to the latest version, which means you might delay your next upgrade just a bit longer.

Join the debate, and respond to today's editorial on the forums

The APPLY function will apply the function specified by fun to the matrix specified by x. In this case, the MARGIN parameter specifies that we look at columns and get the max value for each.

These are 2017, 173, and 40.

Ref: APPLY - click here

In case you're wondering about the numbers, these are Albert Pujols stats with the LAA.

click here

» Discuss this question and answer on the forums