To complement on @Abdul Manaf's answer...
The relational model and SQL won't let you know exactly the cardinality of all your relationships. For instance, and unless you use very tricky techniques I can't think of at the moment, you won't be able to distinguish between a 1:1 relationship from a 1:0...1 relationship.
To have a 1:1 relationship, you would need to define two tables which reference each other for the same values. Although the standard SQL would allow you to do that, most databases will check FK constraings (
REFERENCES) at INSERT/UPDATE time, and not have the test
DEFERRED to the end of the transaction.
This means you are very likely to not have the information within the database to distinguish between a 1:0...1 and a 1:1 relationship. You will have to consider them always 1:0...1.
You normally can distinguish 1:0...1 relationships from 1:0...n relationships, because the column on the left table has got a UNIQUE constraint on it (or is a PRIMARY KEY, which implies a UNIQUE constraint), and the column on the right table has got a UNIQUE constraint (1:0...1 relationship) or does not have it (1:n relationship). Any kind of 1:2 or 1:3 relationship that can be represented in an ER diagram cannot be (for all practical purposes) represented in an SQL schema.
All the same applies, when you change columns with multi-columns, although all the queries to find out get much more complicated; because you have to take into account all columns at once.