WOW!!! The next great update to Execution Plans is a way to watch the processing of rows through iterators while the query runs. Never saw that coming. And even a bonus, you can use SQL Server 2016 Management Studio (SSMS) to watch queries run on versions 2014, 2012 and 2008 R2.
SSMS 2016 can be installed by itself without the server installation. Go here for the installation. Once you install SSMS, you can start to view this addition in SSMS.like Figure 1.
Figure 1 – Live Query Statistics
After opening a New Query, go to the Query menu choice and select “Include Live Query Statistics.” You will need a long running query in order to see the processing happening at a slow pace.
Figure 2 – Include Live Query Statistics
Once you start a query, a new tab will show the Live Query Statistics like Figure 3. The dotted lne means the iterator has not finished or has not started.
Figure 3 – New Tab
The labeling of the iterators have the usually properties as an Execution Plan plus a running time clock of seconds that this iterator is using. In addition, you can see the “Actual Number of Row” of “Esitmated number of Rows” with a calculated Percent Done (Actual/Estimated). Figure 4 shows 144% for some iterators that have not even finished yet.
Figure 4 – Percent Greater Than 100
This is useful to see if estimate number of rows is the same or close to actual number of rows. The percentage will rise above 100 if there are more actual than estimated rows. The query in this case probably could use some performance tuning to get the estimate closer to the actual.
The second half of the day was spent talking with various speakers and attendees. I got into a great conversation about trying to mentor developers at work places about how to write good T-SQL. It was interesting to here how some developers (or DBAs) are not open to new ways to write T-SQL like using CTEs or proper join syntax.
I meet Lance from the program committee and discussed the difficultly of having 800+ submissions and only about 140 slots for those submissions. To make sure you get a wide range of topics and not covering too little or too many of the same topic is not always easy. They do a great job with the volunteers to make sure Summit is a great learning opportunity for all.
I also saw a new posting on SQLServerCentral.Com about learning SQL Server 2016. Try this Link – http://www.sqlservercentral.com/articles/SQL+Server+2016/127518/.
The only session I attended was on learning how to price and scale an Azure SQL Server Data Warehouse. So, now I understand DWUs and pricing is 3 fold: DWUs, Storage and GEO DR. It can be expensive, but you get Massive Parallel Processing in the cloud and easier scaling up or down the usage.
The last 2 session time slots was spent co-presenting with Bill Anton on Performance Tuning Analysis Services. It was well attended but in a large room. There was about 2/3 of the people returning after the break for the 3 hour session. Bill did a great job getting people started with Profiler, PerfMon and some free tools and scripts. Lots of questions and even a Microsoft developer helping us answer questions. Thanks Microsoft.
Last stop of the day was dinner with some great guys from Atlanta at Rock Bottom. They went to a BIML party, I went to bed.
Today’s keynote was by David Dewitt. He did an excellent job explaining the architecture behind various Cloud Data Warehouse offerings. It was nice to get an explanation of the behind the scenes node and disk usage.
One announcement that caught me off guard was next year’s Summit is Oct 31 thru Nov 3. That is Halloween and my kids come first on that day. Not sure that was a good idea. Better information shared was an increase in revenue of 10%, plus 100%+ increase in Global membership in some areas. There is a new logo and the updated website will be released early next year. Denise provided some slides on what the website will look like including mobile friendly.
Next, I am going to meet up with Bill and attend a session on Azure DW.