how should I manage changes in a dimension field when using big flat table?

Let's say that I have a sale (a fact entry) made by a user (a dimension). initially I would insert this sale+user into my big flat table in my data warehouse. So far sounds very standard

But my doubt is about what should I do when a dimension entry have a change?, for example, the user change his address.

How should I proceed in this scenario?, should I introduce a new sale, using the same data as before, with the user information updated, into my big flat table?

NOTE: I'm speaking about dimensions, etc, but I'm not pretending to use at all an star-schema as an intermediate step (at least for my learning problem)

Solution

The user dimension is usually called conformed/shared dimension.

A conformed dimension is a dimension that has exactly the same meaning and content when being referred from different fact tables.

In your case :

But my doubt is about what should I do when a dimension entry have a change?, for example, the user change his address.

If you have a need to store the User's address for analytic purposes, you can opt for Slowly Changing Dimension Type 2 – Row Versioning. Therefore you need to update your fact table and link it to the last version of the User.
If you don't need to track the addresses changes of a User you can use the Slowly Changing Dimension Type 1 – No History. Therefore no need to update the fact rows since they will always point to the unique version of a user