We used a very nice set of controls for this project in Telerik. They are by far my favorite set of proprietary controls for the .NET Framework. The project was heavily built around these controls so it was only natural to want to keep with them. One problem…98% of the charts had to be Logarithmic. If you are not familiar with this…its when a charts Y Axis is equally spaced with labels dramatically increased (1, 10, 100, 1000, so on) instead of your normal Linear charting (1, 2, 3, 4, etc).

This presented a problem because the Telerik RadChart did not represent the logarithmic minor ticks and lines properly. They didn’t increase and scale the way they were suppose to. It’s hard to explain in words, so here is a screenshot:

The image above still shows a Linear scale even though it’s actually Logarithmic. Well, I searched all over the internet to see if this was just a bug in the Telerik RadChart control or if I was just doing something wrong. Turns out, Telerik knows (recently) about this issue and it has been logged. I added a reply to that thread (in the link) asking if there was any update on the issue…and as you can read there…they said they were not going to fix this. They did offer a workaround though…and here is a screenshot:

What they said to do was just remove the Minor Grid Lines and Ticks from the chart. Well, I had already done that as a temporary fix, but if you know anything about the oil and gas industry then you know how sensitive and important the data is. So, when we are charting their data it is very important to be as precise as possible. The minor grid lines were very important. (BTW, this is just test data in the images, so don’t worry about any issues about showing the clients stuff).

I did some more digging and searching because I really didn’t want to spend too much more time and/or money on this Charting need. I was very lucky too. I came across Microsoft’s latest Charting Controls (add-on to Visual Studio 2008). What’s even better is that as of VS 2010 and .NET Framework 4 it will be already incorporated so there won’t be any installation needed. It did take me about a day to configure, get the data bound to it, make it pretty, etc…but this was the solution I ended up using. And, it turned out to be a lot better…and much more visually appealing. Here is the final:

My only other option would have been some Flash solution or Silverlight. And if you know me very well, you know that I HATE Flash and/or Silverlight. This is one time I can say Hurray for Microsoft this time.

I needed to Create a new Cube today, and in the process I figured we needed an ‘UNKNOWN’ row for all those Unknown situations that could pop up in a Pivot table. Well, when you create a Cube, most of your Dimension and Fact tables are driven off of Keys that you set as Identity. Or at least this is how MS came to our shop and showed us how to do it.

Now, the problem here is Identity is READ-ONLY so our staging tables and Views could not properly add a ‘UNKNOWN’ with a regular INSERT when trying to force that Identity value to -1. Why negative one? So that you will ALWAYS know that this is either an UNKNOWN or N/A.

So, here is what we did within the initial load of one of the Dimension tables:

TRUNCATE TABLE PM_DWT.dbo.Dim_Partnership
SET IDENTITY_INSERT PM_DWT.dbo.Dim_Partnership ON
GO
INSERT INTO PM_DWT.dbo.Dim_Partnership (Partnership_Key,Partner_ID,Partnership_Name)
VALUES (-1,-1,'N/A')
GO
SET IDENTITY_INSERT PM_DWT.dbo.Dim_Partnership OFF
GO
DBCC CHECKIDENT (Dim_Partnership, RESEED, 0)
GO
INSERT INTO PM_DWT.dbo.Dim_Partnership (Partner_ID, Partnership_Name)
SELECT
Partner_ID
,Partner_Name
FROM PM_Cluster.dbo.v_Dim_Partnership
ORDER BY Partner_ID
GO

The above code will insert your first record for -1 and UNKNOWN. Then it will enable the Identity column again. What happens here though, is the Identity column see an entry, so it naturally wants to add one, so that your next value inserted will actually have 2. We want to force it to restart even though we inserted a record.

DBCC CHECKIDENT (Dim_Partnership, RESEED, 0)
GO

Will do just that…so that the main INSERT will now begin at 1.

SQL:

SELECT DATEADD(yy, DATEDIFF(yy,0,getdate()), 0),'First Day of the Year'
UNION
SELECT DATEADD(dd,-1,DATEADD(yy,0,DATEADD(yy,DATEDIFF(yy,0,getdate())+1,0))),'Last Day of the Year'


Results:
2008-01-01 00:00:00.000 First Day of the Year
2008-12-31 00:00:00.000 Last Day of the Year