Quote Originally Posted by hunsnowboard
Hi Hansup and CreganTur!
Thank you for your replies. I'll try to answer you both.

Hansup: the table might have duplicate data. But that it is not important. I will delete those rows. Anyway this table is a hierarchy relation table. Actually it represents companies related to each other (by the mean of ownership; this means that the main company is in column Important_0, then there is the other company (column Important_1) which is related (owner, partner, shareholder, etc) to the main company, and then there are the other companies (important_2) which are related (owner, partner, shareholder, etc) to the second company. Now my problem is that there are more then 10.000 companies, so I cannot make a table for each one!

CreganTur: described above you can see my problem more clearly. As for normalization, I read a few sites, I watched a few tutorials but the problem is that everyone shows this customer-order relation, which I understood clearly and already can make it by myself. The problem is that in my case I really do not have a clue how to normalise this table where there are companies related to each other... That is why I am writing on forums, hoping that someone takes a look at my table and gives me a tip/hint how to begin...
I fully agree with Hansup and CreganTur. Normalization is a very broad subject that requires a lot of "learning by doing". There are many reference texts and sites.

When I see fields identified as X-importance1, X_importance2, X_importance3 etc it tends to show that the problem is not yet understood. A big part of normalization is understanding clearly what you are trying to do.

If you start by identifying what things you have and how they are related to identify "your business rules"; then you can create some level of data model (no matter how crude it may be).

You then put your model to the test by taking each of you rules and see if the model supports it.

If not, adjust the model, until you are content that the model supports the rules you have identified.

Then design your tables based on the model. May seem like a long way around, but I assure you that you will understand the problem and the facts you are dealing with much more clearly.

Just some thoughts and observations.

Good luck.

Here are a few questions that I have:

Are you concerned with the relationship-type (owner, partner, shareholder...) ?
How do you differentiate and owner from a shareholder etc?

Can a company only have 1 owner?
How do you deal with multiple owners, if it's important?

What are you really interested in -- what will your data base do/solve?