Question

A coworker sketched out the values of a new table as:

"Foo", "some value 1"
"Foo", "some value 2"
"Foo", "some value 3"
"Bar", "some value 3"

These are the only columns in the table. The column names are Col1, Col2.

One person said that this table is not normalized, another said it is.

The specific argument that it violated normalization is that removing the three records with “Foo” in Col1 “Foo” would no longer be present in the system. That person said there should be a lookup table containing an ID, and Name column. The table above would reference the Id of that table as its FK.

The argument that it wasn’t normalized is that there wasn’t a third column in the table dependent on the first (3rd normalized form).

The confusion I think comes from it being 1NF in that it satisfies this example:

Customer    Tr. ID  Date            Amount
Jones   12890   14-Oct-2003     -87
Jones   12904   15-Oct-2003     -50
Wilkins     12898   14-Oct-2003     -21
Stevens     12907   15-Oct-2003     -18
Stevens     14920   20-Nov-2003     -70
Stevens     15003   27-Nov-2003     -60

from http://en.wikipedia.org/wiki/Database_normalization.

But it sounds like it violates this rule, “The same information can be expressed on multiple rows; therefore updates to the table may result in logical inconsistencies.” This applies to normalization beyond 1NF.

So it looks like the original table would violate 2NF, and thereby 3NF, but would satisfy 1NF. Is this correct?

Answer 1

If those two columns are really all there are, then I would say this database table is in third normal form. Here’s my reasoning:

1. It’s CLEARLY in 1NF since none of the attributes are “multi-valued”
2. Since neither col1 nor col2 are a valid key candidate (duplicate values!), the only possible and valid primary key on this table is (col1,col2)
3. 2NF stipulates that no non-prime attribute shall be functionally dependent on a part of a candidate key. Since there are only col1 and col2 which are both part of the only possible candidate key, this point is moot – the table IS in 2NF
4. 3NF according to E.F.Codd basically says that any non-key attribute must be dependent “on the key, the whole key, and nothing but the key”. Since we ONLY have two columns which make up the key, there are no other non-key attributes, so none of the non-key attributes violates this rule –> the table IS is 3NF

I don’t know if your work buddy wants to really get into 4NF, 5NF or Boyce-Codd NF – I highly doubt it……

Marc