icCube mapping a dimension attribute to a measure group - iccube

I'm trying to figure out the best way to link a dimension to a fact table and having some trouble finding an example in the documentation. All of my data sources are csv files and I have the following 2 data tables:
Areas table with columns: Area,DateTime,Cost (Area,DateTime are unique)
Company table with columns: CompanyID,Area (CompanyID is unique and represents a dimension)
I would like to link the Company Dimension to the Cost measure from the Areas table. However, it seems that I can only link the Company dimension to the Areas measure group through the dimension key, which is CompanyID. Is there a way around this or do I need to add a CompanyID column to my Areas csv file prior to loading.
thanks

I'm not really sure what you want to achieve but it looks as you could use a many-to-many cube to create a link from Area to Company (even though it might be a one-to-on relation).
1) Create a many-to-many cube using your Company table (Advanced/Many-to-Many)
2) Bind the Company dimension in the facts, Area table, using the defined many-to-many relation.
Some documentation here (the first image is wrong).
hope it helps.

Related

Can the same dimension table be related to multiple fact tables?

I am new to OLAP,if I have two fact tables can they share the same dimension table?
A good example would be if I have tables fact1 and fact2, can they both have a foreign key into a single Date dimension (dimDate) table? Or, do I need/should create separate dimDate dimension tables for each separate fact?
To me, and based on my research, I don't see any downfall of sharing a dim table, but wanted to check.
Thanks!
They can, and should.
That's the whole point of conformed dimensions, keeping the attributes in a single place, so as to avoid multiple versions of truth coming from different fact tables.
So a single date dimension, with all the necessary attributes for each fact table, which is then linked from each fact table that needs it.
Same for a customer dimension. If you have a sales fact table that needs customer info such as billing address and a marketing fact table that holds info about campaigns each customer can benefit from, you would combine all those attributes in a single table. Some customers may not be referenced in the marketing fact table, others may not exist in the fact table, but all would exist in the single customer dimension, which is your single source of truth about who your customers are.

Tableau---Getting count from 2 different data sources and combining into one total

I am a tableau newbie and am trying to see if this is possible or not. I have 2 separate data sources where the same employees are listed, one is for closed cases and the other is for open cases. These data sources have some of the same columns, but for the most part they are different.
Is it possible to aggregate the case count for each employee on the closed and open data sources into a single column? For instance, if an employee has 50 closed cases and 23 open cases, I want it to show 73 for them.
I attempted to play around with the joins/unions but these didn't work properly and duplicated the data most times.
I think this is a great chance to leverage blends.
I have created a workbook with the Sample Superstore Excel dataset. This dataset has three sheets. I'll use the Orders and Returns sheets to demonstrate how we can calculate the net orders using blends.
The dataset I'm using can be found here.
Start by connecting to both the Orders and Returns separately. Once done with this step you should see the two data sources at the top of your data pane.
In this example, I'll calculate the Net Returns by Category. In your case, you're after the Total Cases by Employee, so just imagine Employee in place of Category.
Next, drag Category from the Orders data source onto the view, then select the Orders data source and click the chain icon to blend on Order ID.
You will need a common column between the two tables in order to blend.
Once blended I'll go back to the primary data source (indicated by the blue check mark) and create the Net Orders calculation.
This calculation uses the dot notation - similar to what you might see in SQL - to reference our other table.
To double check that our calculation is working properly, we can drag the components of this calculation onto the view and do the math.
Of course, once you are satisfied you can remove all but your blended calculation.
Blending isn't ideal in most cases but you could try it. Bring in each data source separately and "join" them within your workbook pane on Employee or hopefully an Employee_id. Click the little chain once you have them both loaded and you are on a worksheet tab. Then you could sum the counts by employee. Blending sometimes presents some issues with calculated fields across the two data sources but this is what I would try first.

ER diagram - design issues

There are 3 entities:
vehicle_model
vehicle
extra_options (such as open top, leather seats, etc..)
Vehicle model can have a subset of the extra options.
Vehicle can have a subset of it's model extras.
I've been trying hours to figure out how to represent this as er diagram, but without success. I Thought about ternary relationship ,and although I don't understand it completely I think this isn't the way.
I thought about creating another 2 entities, model_ext & vehicle_ext ,so that vehicle_ext would be connected to model_ext but this isn't a good design.
This is my first er diagram design. I'm really lost (read er-diagram chapter in "Silberschatz, Database System Concepts" three times already) so any idea would be appreciated.
did you try adding a new table say 'vehicle_vehicle_model_extra_options_map'? (you can name this table to any thing short, but for better explanation i use __map as a standard way for defining the map tables.)
note those two null able foreign key columns in this table.
Basically, vehicle has one to many relation to extra_options, and vehicle_model has one to many relation to extra_options table, therefore the new table was added.
updated:

Do we have to use fact table for reports?

I am working on building a data mart for reporting purpose.
I am new to this field and looking for help.
I have a fact table and two dimension tables.
The fact table has only 3 fields, its primary key and foreign key references to two dimension tables.
The two dimension tables have data related to 1)phonenumbers and 2)extension numbers.
(I cannot combine these dimension tables because they have different information)
As you see my fact table does not have any quantitative columns.
I want to generate a report that displays phonenumbers and corresponding extensions.
I can get this information by performing a join on the two dimension tables.
So my question is do I have to use fact table for the report? i.e Should I first get the key from phonenumber table, perform join on fact table, get extension key and perform join on extension table?
OR
Simply join the two dimension tables to generate the report because it is possible in this case?
Do we have to involve the fact table?
Thanks for reading.
Any help is appreciated.
do I have to use fact table for the report? i.e Should I first get the key from phonenumber table, perform join on fact table, get extension key and perform join on extension table?
Often, this is necessary.
Simply join the two dimension tables to generate the report because it is possible in this case?
Sometimes, this works, also.
Do we have to involve the fact table?
Depends on the relationships.
If you have a "hierarchy" of dimensional information, then the two dimensions could be directly related. In this case, the fact table doesn't tie them together. The fact ties to the detailed dimension; the detailed dimension ties to the summary. This is rare.
Dimensions change.
If you have two or more Slowly Changing Dimensions, then your dimensions may include lots of "previous" relationship information.
Fact 1: Phone xxx-xxx-xxxx, Extension yyyy
Fact 2: Phone xxx-xxx-xxxx, Extension zzzz
Then, another load applies an SCD rule to modify zzzz to aaaa, as of 7/1/11 You may have the old dimension values available, as well as the new dimension values, with an applicable date range.
Now, the fact (and the date range) are required to define which copy of the dimension value you're going to get.
Fact 2: Phone xxx-xxx-xxxx, Extension zzzz, from beginning to before 7/1/11.
Fact 2: Phone xxx-xxx-xxxx, Extension aaaa, from 7/1/11 to end.
So, you may need the fact, dimensions and time to sort out the relationships.

Cube Design - Bridge Tables for Many To Many mapping with additional column

Am making a cube in SQL Server Analysis Services 2005 and have a question about many to many relationships.
I have a many to many relationship between two entities that contains an additional descriptive column as part of the relationship.
I understand that I may need to a bridge table to model the relationship but I am not sure where to store the additional column - in the bridge table or elsewhere?
Many To Many relationsip in SSAS can be implemented via an intermediate fact table that contains both dimension key that subject to the relation.
For example; If you have a cube that has a book-sales-fact table and you want to aggregate total sales by author (which may have many books and a book may be written by many authors) you should also have a author-book intermediate fact table (just like in relational database world). In this bridge table, you should have both dimension keys (Author and Book) plus some measure related to the current book and author such as wages paid to the author to write the book (or chapters).
As a result, if your additional column is kind of a measure you should add that column to the intermediate fact table.

Resources