Miracle is a sponsor for SQL Saturday Iceland and therefore I am posting this on behalf of the preparation team.
The second annual SQL Saturday Iceland free training event is approaching fast. On March 18th around 20 speakers will be in Iceland to tell us all about the Microsoft Data Platform. There will be three tracks to choose from: one focusing on developers, one focusing on DBAs and one focusing on BI. There are 21 sessions so everybody should be able to find something to their liking. The conference is held at Reykjavik University and the sessions are in English. The best thing about this is that access is completely free and the conference is open to all. Although the sessions are free we ask that you register as we will also be serving complimentary lunch and we need to estimate the required quantities. Please register at https://www.sqlsaturday.com/602/RegisterNow.aspx.
Recently, I read an article titled ‘The Myth of Self-Service Analytics’ (http://www.perceptualedge.com/blog/?p=2467) which discusses the danger of jumping on magic solutions based on colourful software tools which, supposedly, can enable everyone and anyone to analyse information and business data. My favourite gripe is about Self-service BI because there is more than meets the eye when it comes to this topic. There are several points to keep in mind while discussing Self-service BI because they are, as usual, not black and white but grey.
- BI companies want to sell their Self-service BI (SSBI) tools as widely as possible.
- The target group for SSBI tools is the management level. Data analysts and information professionals are preferably left out of the sales process as they understand what is really needed to make sense of data and present information in an appropriate manner.
- Self-service BI can NEVER work without a data warehouse of some kind. This means that the whole concept is, to a point, a deception. “Here is your tool, here is your database with 25 thousand tables. Give me sales figures and a forecast for the next year tomorrow.” It goes without saying that this is an unrealistic scenario.
- SSBI can work for superusers who work with a set of well-defined data. Nevertheless, there is always a risk of more than one version of the truth under those circumstances. Of course, that danger is also present when data scientists and information professionals are involved but the risk is greatly reduced.
- SSBI has been around for a long time but not necessarily under that name. DSS – DW – OLAP – Analytics – SSBI. These are not necessarily exactly the same concepts but there is a very big overlap. To a certain extent making users self-sufficient has always been the goal of BI and it is certainly possible but it doesn’t happen without a solid foundation and that foundation is your data warehouse.
I came across an article that describes a mechanism in Oracle that automatically performs a function that has traditionally been implemented “manually”. It caches results from functions so they only have to be called once for each value sent to the function. This works for version 11.1 and newer and version 11.2 offers some improvements. The benefit of the way Oracle implements this is that the mechanism works across sessions, i.e. if function f(x) is called in session 1 then the function does not have to be called again in session 2 because the result is cached if RESULT_CACHE is correctly configured.
The article offers a very good explanation of these concepts: https://oracle-base.com/articles/11g/cross-session-plsql-function-result-cache-11gr1.