Redundancy of information and algorithm to do lossless decompositions

dependencies

Let be a relation R(NameFile, Size, Directory, DateCreation,HourCreation,Login,AutAccess,DateAccess,TypeAcces

F={NF,D→S;
   NF, D →DC;
   NF, D→HC;
   L,NF,D→A;
   L,NF,D,DA,HA→TA;
}

Part of the instances are given in the following array:

Example

  • Does the relation create information redundancy?
  • What would a decomposition of R without loss of information and without loss of dependency be?

I don't think so because it seems that the information given by the dependencies seem to be unique.

I know there is algorithm to decompose R without loss of information or without loss of dependency but I don't know them

Best Answer

Does the relation create information redundancy?

Yes, for instance you have the information on the file (name, size, data creation, etc.) repeated for different accesses to the same file.

What would a decomposition of R without loss of information and without loss of dependency be?

The following decomposition is both in Third Normal Form and in Boyce-Codd Normal Form and is without loss of data and loss of dependencies. It can be obtained by applying both the “analysis algorithm” to produce the BCNF, and the “synthesis algorithm” to produce the 3NF, but note that in general only the latter is guaranteed to avoid any loss of dependencies:

R1 < (D DC HC NF S),
{ D NF → DC
D NF → HC
D NF → S } >

R2 < (A D L NF),
{ D L NF → A } >

R3 < (D DA HA L NF TA),
{ D DA HA L NF → TA } >

Any good database book will describe both the algorithms.

Related Question