Column name collision in SQL Server view joining database architecture - sql-server-2008

Question

First of all, sorry the way I write, bad English here.
Please, consider the following scenario:
I made my database architecture combining views as such:
Entity City has the columns { city_id, city_name, state_name, ... }
Entity User has the columns { user_id, user_name, user_login, city_id, ... }
Each entity has a view, the city view does not have FK to any other entity, so its a simple select.
User entity have one FK to city entity, so it makes a join with the city entity in the view.
As the following example:
create view vw_user as
select user.user_id,
user.user_name,
user.user_login,
vw_city.*
from user (nolock)
inner join vw_city (nolock) on vw_city.city_id = user.city_id
go
So if you have an entity like user_access with a FK to user, the view of user_access will have an inner join to vw_user, and it will bring up all the columns from user and city entities.
In the end, I just make a query on the view, and it returns the full entity with all the foreign references.
This works great, and make the maintenance procedures very easier, I have a stored procedure that recompile all the views if any change is needed to one of the entities.
But, it has a problem, and I discovered this in a very bad situation, the system that uses this architecture is working very fast and smooth already, and I need to join two entities that share the entity city for example, so it generates a collision of columns, SQL Server does not allow this in views, you can execute a query with collision with no problem, but you can't do it in a view.
So, this is my problem, I need to find a way to fix this, but I can't find an answer to it myself.
The only thing I ended up doing is making a hard-coded view with the columns renamed.
But I want a solution for the entire architecture, something permanent, like an upgrade.
This architecture is followed in code, with C#, so there is a dependence on the column names.
When I read the data from the query, I need to get the data from the column name, so this code can be reused with other classes, doing the same thing that the view does in the query, in the c# code.
So, getting data with ordinal is out of the question.
Any ideas?
Thanks, and sorry, but this is very hard to describe.

Answer 1

Well, just to close the subject, as SMC suggested i have used aliases to identify the columns and the tables in my queries.
So the query looks like this now - using the example from the topic:
Select 1.user_id as 1,
1.user_name as 2,
1.user_login as 3,
1.city_id as 4,
2.city_id as 5,
2.city_name as 6,
2.state_name as 7
from user as 1
inner join city as 2 on 2.city_id = 1.city_id
Just notice the columns 4 and 5, they will not collide now :)
Wen my application makes this query he also make the Classes involved to know their aliases in the current query.
It works fine, solved my problem and i don't need to use views anymore.
In the end, the application is faster then before.
That's it :) hope it helps anybody out there.

Answer 2

It's because the Column Name in each View or Function must be unique. You need to be specific while selecting the column Names. Use Alias and then refer the column Names.
You tried to create VIEW using a single column name more than once in the statement.
Remove * and give explicit column name