Although SQL Server 2005 RTM is around for more than half a year and has already service pack 1, I still get many questions about the new features it brings. People try to realize why should they upgrade their current SQL Server 2000 installations or why should they migrate from other systems. Therefore, I decided to write couple of blogs with brief explanation of major novelties together with an evaluation of usefulness of them. I have to warn you: everything I write here is strictly my personal point of view. Still, I am going to corroborate my opinions with reasons why I think this way. In addition, I am not going to talk about cool features exclusively; I intend to mention what I dislike as well.
In the first place, I have to mention the Data Mining part. I am impressed how good the whole thing is. For the first time, data mining is available for masses. It is not just because data mining is for free if you have a legal SQL Server 2005 copy; the important thing is how easy it is to use as well. I think Microsoft is going to shake the data mining market in next few years the same way as it has shaken the OLAP market with SQL Server 7.0 & 2000. The complete process of data mining is integrated with other parts of the suite. You can mine relational and OLAP data. There are many extraordinary model viewers integrated in BI Dev Studio. You can use a data mining model as a source for Reporting Services. You can deploy a model as a dimension in an OLAP cube with just couple of clicks. You can use a model for smart data cleansing inside your SSIS package, where you get Data Mining Task in the control flow and Data Mining Transformation in the data flow. Of course, SSIS supports inverse direction as well – there are many tasks and transformations suited for data preparation for data mining. SSIS still supports the AS Processing task known from DTS 2000, which you can use to train a model. In addition, if you want to include data mining models in your code to make your application smarter, you get ADOMD.NET and DMX language. If you want to present the model graphically in you application, you get Windows & Web form viewers. Finally, if you wan to create and manage models programmatically using AMO.
Microsoft is not very famous among data mining specialists yet. After all, MS is a new player in this market, if you leave out of account data mining in AS 2000, which was IMO more or less for trial purposes only. However, the way they did it in SQL Server 2005 gives me assurance MS is going to become a leader in the market very soon. In short, you need only the data mining knowledge; the integration and the tools are already there. I am even more confident MS has done a right thing again because I have seen the Magic Quadrant for Customer Data Mining for the first quarter of 2006 by Gartner. Microsoft is not even mentioned there. I remember something similar for OLAP in AS 7.0 & 2000: Gartner underestimated MS as well, and now MS is a market leader. It seems to me Gartner is too much focused on existing major players and overlooks the advantages of SQL Server 2005, namely integration, tools and price.
- Python for SQL Server Specialists Part 4: Python and SQL Server - April 24, 2018
- Python for SQL Server Specialists Part 3: Graphs and Machine Learning - April 11, 2018
- Python for SQL Server Specialists Part 2: Working with Data - March 22, 2018