August 2008 - Posts
My favorite "toy" in SQL Server 2008 just got a brand new spanking release! Report Builder 2.0 just went RC1 after it didn't appear in the feature pack of SQL Server 2008 RTM. When I tried it out first, as a Report Designer Preview in CTP6, I loved it so much that I even decided to make a webcast out of it (both in Hebrew and in English). Later, I found out this would become the "next generation" of Report Builder. When I downloaded the feature pack that came with SQL Server 2008 RC0 I admit I was a bit disappointed, as they took out the link directing me to the Data Source definition, and I for one really liked that functionality. But it all seems to have been worth the wait. First of all, they have changed the first screen again and what you see will start you on your way to either a Tablix or a chart (though obviously you can add later one of the other):
And then you start your way on a wizard, going through from creating the Data Source:
Creating a query with joins and a start of a parameter with just a few clicks:
Creating groups with a simple drag and drop:
And formatting those groups is just as easy:
After a bit of styling, all that's left is to run the report:
If you remember, in Report Builder 2.0 CTP6 you didn't have a wizard to help you build a query and a parameter so easily. I think I may soon find myself becoming obsolete for quite a few of my users... And that's not all because you can also open and edit Data Sources and Reports saved on the Report Server.
So, if you're excited just as much as I am, I strongly recommend you download Report Builder 2.0 in its RC1 version. The final release of Report Builder 2.0 is scheduled for fall of this year, along with the updated feature pack for SQL Server 2008.
CU9 for SQL Server 2005 Service Pack 2 has just been released, which I should have guessed as the grounds have also been made for Cumulative Update Package 10 as well.
It seems that you may need CU9 to correct a bug caused from installing Cumulative Update Package 7. What's the problem? Apparently you may get an error message after installing CU7, when you explicitly define a slice at a partition and then process the partition by using a query that returns no records. To read more about the problem that may arise from CU7, that was fixed in CU9, please check the Knowledge Base Article detailing the error message that can come when you process a partition. Another problem that's fixed is for an MDX query that's run against a Parent - Child dimension that has the HideMemberIf property set to ParentIsBlankSelfOrMissing.
You can read further about Cumulative Update Package 9 for SQL Server 2005 SP 2 and then also request a download of CU9.
Still, I have to ask, now that SQL Server 2008 has RTM-ed, when is SP3 going to be released?...
Today, I understood the difference between reading about a car accident, and witnessing one.
I was on my way to work this morning and I had crossed the street to the other side. I even managed to take a few steps towards the office. Suddenly, I heard a screech of breaks from behind me. I turned around to see on the other side of the street, a man flying off in the air and land on the asphalt. I screamed.
After a second, I noticed the accident involved two people, the driver of the mini - bike (which I saw getting thrown off his mini - bike) and a pedestrian who had started crossing the street. People all around stopped and rushed to their aid. The driver of the mini - bike now had his helmet off. He was sitting on the asphalt shaking his head in disbelief, as a man who came to his aid was putting his hand on his shoulder. The pedestrian wasn't as lucky. He couldn't get up. A man was holding his hand and rubbing it while another supported his back, as he was lying on the side. Someone else was calling an ambulance, and all I could think was how lucky we're so close to the hospital.
An ambulance came within a few minutes. The driver was walking around anxiously. The pedestrian stayed lying on the asphalt. The ambulance took the pedestrian to the hospital, as police started questioning the driver.
There are so many accidents happening nowadays. Please think of your loved ones at home. Don't become part of the statistics, drive safely.
I know I had recommended the BIDS Helper for you before, but as a new version of this Visual Studio add – in just got out this Saturday with support for SQL Server 2008, I thought it's a great time to remind everybody they should check it out.
Most of the features in this release are available only through their source code, which is there for you to compile and deploy yourself. My personal favorite feature would be the Non-Default Properties Report, which helps you track the settings in your SSIS Package or your SSAS cube and dimensions for varying properties. This should really help you monitor all the special definitions you made in your cube, without having to search all those property dialog boxes (which would be somewhat tiring, to say the least).
If you've been following the new features for Analysis Services 2008, then you probably recognize quite a few of them from the BIDS Helper which was available for you already from Analysis Services 2005. So now that there's a release for BIDS Helper for SQL Server 2008, don't you want to see what will be incorporated in the next version of SQL Server as well?...
Till now Microsoft enabled you to download SQL Server 2008 for a trial version of 180 days. You could get it from MSDN for the Developers and Architects of you and from TechNet for IT Professionals and IT Managers (you get the same download on both sites, only with differently focused follow – up emails, according to your role).
Yesterday however, Microsoft also released SQL Server 2008 Express. Yes, I grant you, it does not have full functionality in it like SQL Server 2008, but I personally would be very happy to play around with what’s already inside and not think about 180 days till the expiration of the trial software. It has PowerShell Integration and supports new Date & Time data types, but at the same time, it still doesn’t yet include, amongst other things, the enhanced Gauges & Charting which were added to Reporting Services 2008 or the ability to export your reports to Word. For those, we’ll have to wait a bit longer till SQL Server 2008 Express with Advanced Services comes out, around the end of August.
So if you’re ready to start playing – go to download SQL Server 2008 Express.
Yes, it’s finally out – THE NEW VERSION of SqlServer , 2008 ;)
I found myself revisiting the Parent Child dimension quite unexpectedly.
I'm working on a project which involves the division of the city to different statistical areas. The division of the city into these different statistical areas happens once every 10 years or so, according to the changes in population tracked by the municipality itself and the government in the Census of Population and Housing.
Up till now, my cube had in it just the statistical areas distribution from one Census onwards and so had just one sort of distribution I needed to connect to in my cube. However, now I was asked to add historical data which included in it the distribution that existed before the current Census. The current distribution of the city into areas is either similar to the previous version, or at times a split of a certain areas into two new areas (for instance). If it stopped there, we could say that the lowest granularity level I need to connect to in the fact tables in my SSAS project are always those of the last Census. Unfortunately enough, that isn't the case. I was also handed a draft of the statistical areas forecasted for the Census in one year's time. This time around, there were areas which were some times identical, sometimes a merger of other areas and at other times a division of other areas. So, in effect, I didn't have anymore a certain Census for which the areas in it consisted for the lowest granularity. I needed to find a new way to define a granularity level for this dimension in my cube. I guess from the title you can infer the solution…
My statistical area is now no longer the simple hierarchy it used to be. Instead, I found it will be a different hierarchy for each Census conducted. Also, no hierarchy is necessarily of a lower granularity than another. In the new dimension – which is a Parent Child dimension - the areas would now be defined not just by their ID number, but also by the Census year to which they belong. The Parent Child dimension will now also have a fictitious level for each division or unification that happened. Considering that each area consists of 3 digits, and that I have areas before the Census of 1995, after it and after that of 2009, my Parent Child dimension could look somewhat like this:
While looking at the above, we have to remember that a new member can be created at each level of the Parent Child Dimension. I no longer have a balanced hierarchy, as each Census may lead to a new member created based on, or stemming from, one of the existing children. This is due to the fact that each area may become a parent to two or more areas (the lower branch of the tree) and a new area may become the parent of two or more merged areas (the upper branch of the tree).
Before, my Parent Child Dimension came from the fact I needed to connect to data at non – leaf levels, but I knew I had 4 levels in my organizational hierarchy, now I'm not sure how many levels I have in my Parent Child Dimension. This may mean that I may not be able to manipulate my Dimension to show in Report Builder.
Now all that’s left is to redefine the keys in the fact tables in my cube, do checks to see it was done properly, redefine the Panorama reports…