WhenI have a table with 2 attributes only and without dependenceno dependencies. Can I split them as below? Is it in 1NF? Thanks in advance!!!! Original table:
Staff,Skill
A,Cleaning
A,Repairing
A,Coloring
B,Repairing
B,Cleaning
B,Polishing
Split into:
[Staffs]
Staffs (PK)
A
B
[Skills]
skill (PK)
Cleaning
Coloring
Repairing
Polishing
[Staff_Skill] (bridge)
Staff(PK)(FK),Skill(PK)(FK)
Staff,Skill
A,Cleaning
A,Repairing
A,Coloring
B,Repairing
B,Cleaning
B,Polishing
Edit version:
Actually my tables are as below. Are they logical? Thanks in advance!!!
Tables
**Edit 2 **
I try my best to explain clearly
Table Question 2
Normalization replaces a relation by other ones that join back to it. This is to get rid of problematic FDs (functional dependencies) and JDs (join dependencies). If a relation, like this one, is not subject to any non-trivial FDs then it is in BCNF and normalization theory doesn't suggest decomposition. This relation also isn't subject to any non-trivial JDs (it isn't the join of any set of projections of it unless that set contains the original relation), so it is in 5NF and normalization theory doesn't suggest any decomposition. On the other hand, you can decompose it into those three components/projections (per a certain trivial JD) because they do always join back to the original--unhelpfully, since one of them is the original. Since you have decomposed it (and not distinguished between "no FDs" and "no trivial FDs", and not distinguished between "should" vs "can" decompose, and not distinguished between FDs & JDs, and not clarified "normalize", and you don't understand "1NF"), you need to learn what normalization is, from a college/university textbook/course.
PS If you want to introduce Staffs & Skills so that you can record that there are staff or skills regardless of whether there are any current pairings of those, then that is not suggested by normalization but by your noticing that your original 1-table design can't record everything you are interested in. (In that case the two new tables are not always projections of the original, so they can't have been introduced by normalization.)
