“This looks like dense, hmm, may be sparse …I am not sure”
After this topic of my blog, I’ll try to reduce the number of people questioning this one.
I have always been told by many that I explain things from basics. The same goes here too.
First, we’ll spend some time on dimensional model.
No, don’t close the window right now .I mean, only 2 terms of dimensional model.
I shall talk of
Ex: The sales of Honda cars, in Charlotte of North Carolina, in the month of Jan of year 2008 are 5000.
Request all to re look at the above sentence and pull out the dimension names
In the first go, I can make out the following
More scrutiny would give more details
Now, what are these numbers or names suggest .Lets try with the names of dimensions
1. Products (from Honda)
2. Place (from charlotte)
3. Time (from Jan)
4. Sales (let it be this way, for some time)
5. Time (from 2008 year)
6. 5000 (this is the number, which is of sales)
Now, let’s refine it further.
Lets assume that –
There might be more products i.e. Honda, GM, merc
Place, In US, North Carolina is a state and charlotte is one of the cities
Time, this one has a year , quarter , month ..Etc
Now, the final outline can be made with the limited knowledge which we had gained from the above
-Sales (Lets keep is the same , for some more time )
This looks near to our outline.
Fact table contents are the ones, which a user/analyst/decision maker is interested to look at, and understand the business of his company.
The fact table content, when seen against other dimensions ( in our case , time , place , product) gives a user more information about his company.
I.e. Sales in a place, in the month of XYZ for a product.
Now, we are clear with FACT and DIMENSION and what fills them in a dimensional model
Coming to the DENSE, SPARSE, this is our focus
Generally, Dense has contents of fact table and sparse has the rest of dimensions.
To elaborate, Sales is the one, which is a measure/ metric which a typical user would be interested to know of his company.
If you had looked at the definition given in the DBAG (our Bhagvathgeetha) , it says that it has the maximum probability of occurance .
Now ,its easy to interpret that , we might ask questions like
Sales of Honda in Jan
Sales of GM in New jersey
Sales of Honda in 2000
We would be posing questions like
Honda in jan in New Jersey…? And what …can’t even make a question from user’s perspective
Thus, ‘sales’ as a dimension member has max probability of occurrence.
Hence, it’s a dense member and other dimensions like Place, Time , Product are sparse .
I shall continue this post with topic
DENSE OR SPARSE 2..
Questions feed back , I invite
I invite blog topics too ,as this topic is also made on demand and need basis.
Hope it helps