SQL Saturday Houston is now in the books (http://www.sqlsaturday.com/107/eventhome.aspx). This was a great time to meet people from that other large Texas city. I was much more relaxed about my presentation this time around so I was better able to focus on the other presentations.
I started my day with Kendal Van Dyke's introduction to XML (http://www.sqlsaturday.com/viewsession.aspx?sat=107&sessionid=8551). He gave a good summary of how XML is used in SQL Server. For me the highlight was his demonstration of how XML indexes affect the query plans. I understand better the importance of these indexes if I need to query XML columns.
Rick Morelan's presentation about the counterintuitive way that SQL Server handles errors was unsettling (http://www.sqlsaturday.com/viewsession.aspx?sat=107&sessionid=8075). He displayed two stored procedures, the first with multiple steps. Most of us believed that errors on any of the steps would cause the batch to stop after severe errors occurred. However, the batch continued happily on to finish even after encountering these errors. After seeing this, I realized that I am going to need to test my stored procedure code much better to make sure that the errors act exactly the way I expect them to and to apply his suggestions for fixing the problems.
William Pearson III's presentation about using many-to-many dimensions with multiple calendars was interesting (http://www.sqlsaturday.com/viewsession.aspx?sat=107&sessionid=6070). There are multiple calendars that one can relate the facts to. Obviously there is the standard Gregorian calendar. But there are also fiscal calendars. The sample Adventureworks approach to this problem is to add multiple attribute hierarchies to the date dimension (btw, I am a beginning SSAS person, so I may have messed this terminology up). He suggested using many-to-many dimensions, moving the calendar dimension to a separate dimension. He showed how he was able to build this, while at the same time providing the ability for the business uses to prove that this approach worked
After the very excellent BBQ lunch (if it is Texas or Kansas City, it has to be BBQ, right?), I presented my introduction to using the XML Source in SSIS (http://www.sqlsaturday.com/viewsession.aspx?sat=107&sessionid=8531).
I then went to hear Kevin Boles present about T-SQL changes in 2012 (http://www.sqlsaturday.com/viewsession.aspx?sat=107&sessionid=7720). There are several new CLR functions available to use, if you dare. They can affect performance on the box, so be careful. On the other hand there is the Sequence object (http://www.bidn.com/blogs/marcoadf/bidn-blog/2396/sql-server-2012-sequence-object), that as far as I can tell does not affect performance. This object which acts very much like an Identify column on steroids, generates a sequence of numbers that I can use as identifiers for one or more tables. I can see many uses for this, which I may discuss later. The other hot feature of 2012 T-SQL is the ability to rethrow an exception in structured error handling (http://msdn.microsoft.com/en-us/library/ee677615.aspx). This will considerably simplify the code that I have written for error handling.
The final session of the day for me was Bryan Smith's presentation about Big Data (http://www.sqlsaturday.com/viewsession.aspx?sat=107&sessionid=6112). There is an incredible increase in data stored on computers, the vast majority of which it is unstructured. The challenge is gathering business information from that mound of unstructured data, many times seconds after it hits the internet. I can see why this is a very hot topic.
It was a very profitable day for me. My thanks to Nancy Wilson and her team for hosting this event.