Today is Saturday and as such the second day of the Islamic weekend. We however were at work until 7 pm. These 11 hour days get to you since I also had a powerpoint presentation to do, my laundry and other stuff. I must say though that doing laundry in a hotel only involves sticking the dirty stuff in a bag and having it picked up. I am a little concerned though since a very strange guy picked up by laundry (he was no more than 5 feet tall, he was „cirkus sized“) and then called me about 2 hours later informing me that some of my underwear had been damaged....I‘m eagerly awaiting the return of my underwear. News on my underwear tomorrow.
A factoid. 288 people were killed here in car accidents last year. This is a nation of 900.000 people. This is a huge number and I‘m not surprised I have seldom or ever seen worse drivers than here. Parking is like shopping at a sale. Buy 2, get third for free. Except that this means that almost always parking is done in 1 and a half parking spot. Not sometimes or often. Almost always.
An interesting incident at work. A mouse appeared in the hallway right in front of the server room. I‘ve seen a lot of cockroaches, but so far no rodents. This particular mouse was completely calm so we could walk over to it and take a look from 1 a distance of maybe 1-2 feet. As we were looking at it, it suddenly fell on the side and after maybe 10 seconds it rose again and took off– then we found the explanation, it was a Microsoft mouse. It just needed re-booting.
But back to work. I have now completed populating 7 indicators out of the 38 planned for this time. I‘m pretty optimistic that I‘ll finish in time. There has been a lot of time spent on designing and redesigning fact and dimension layout and tomorrow morning I have a little presentation of the design for the purpose of knowledge transfer. Due to the design, the Discoverer reports become amazingly simple. The bottom line is that you always join to all the dimensions involved, just at different levels. So if I wanted the total number for all females in all countries and all cities and of all nationalities I would only have to select 1 record, albeit with the correct conditions. As a result, there simply is no waiting. Unless waiting for a page to re-draw itself. And the data is sparser than we thought so the number of records is going nowhere near the initial estimate of 100.000.000 records. If this were a traditional multi-dimensional cube such as Microsoft Analysis Services or Oracle OLAP these cubes would be very big and might be slower too. Multidimensional database will not handle sparsity as well as relational tables. The drawback on the relational design is that dealing with things such as timelines and complex calculations can be painful. I have to say though that Oracle‘s Analytical Functions are a lifesaver and leave SQL Server in the dust. I might post something on the MODEL clause later which I find very interesting, even though I still haven‘t used it in production. I really have to find a project where I can use it....