0

Let's say I have a Relation R := {id, name, emails} Obviously id -> name is a functional dependency but is id -> emails also a Functional Dependency?

From what I understand id ->> email would be a multivalued dependency if you would turn the Relation into R := {id, name, email}.

But does this only apply after normalization and would it be functionally dependent before?

Scitur
  • 31
  • 1
  • 5
  • Functional dependences and in general all the normalization theory is defined in the relational data model. If an attribute is not atomic then its “container” is a generic table, not a “relation” as intended in the normalization theory, and for this reason the concepts and definitions of the theory cannot be applied to it. – Renzo Apr 15 '21 at 09:55
  • @Renzo Normalization theory can be devided into 2 orthogoanal notions, 1NF & higher NFs. Other than possibly requiring a 1NF, higher NFs don't care about the types of attributes, which can be anything, including relations. A relation by definition has 1 value per attribute per tuple. There is no problem having a FD determining a relation-valued attribute. – philipxy Apr 15 '21 at 12:07
  • What is your 1 (specific researched non-duplicate) question? [ask] Where & how are you 1st stuck finding, quoting & applying definitions of "relation", "FD", "MVD, "normalization" & "atomic"? (From an authoritative source, eg a (good) textbook.) What do you mean, "atomic FD"? "Normalized" has 2 senses, "... to 1NF" & "... to higher NFs".--What do you mean by it? Why is id -> name "obviously" a FD? Why isn't id -> emails? PS By definition a relation has 1 value per attribute per tuple. ["1NF" has many meanings. "Atomic" means nothing in particular.](https://stackoverflow.com/a/40640962/3404097) – philipxy Apr 15 '21 at 12:14

0 Answers0