Original: Reprinted indicate the original address https://www.cnblogs.com/fanerwei222/p/11706874.html
Relational database design paradigm.
1️⃣ first paradigm: (basic NF) can no longer split each column.
Example: "height and weight" column, but also split into "height" and "weight" two!
2️⃣ second paradigm: (on the basis of 1NF) entirely dependent on non-primary key column of the primary key. (To eliminate non-primary key column basis 1NF part dependent on the primary key (primary key) in)
Examples: "Order Number" column and "product number" as the primary key of the row to be displayed the information specific item corresponding to the corresponding item number order number is, assuming that the "order value" column, the "order value" column will only depend on the "order number", belong part dependent on the primary key, this situation can be resolved sub-table, the "order value" column into a secondary table.
3️⃣ third paradigm: (2NF on basis) of non-primary key column does not depend on other non-primary key column. (To eliminate non-primary key column of the transfer on the basis of the dependence 2NF)
Examples: primary key column and a direct correlation, rather than indirectly associated. "Order Number" column primary key "single user id", "single user name", "Single-User gender" non-primary key column, "Single-User name" and "Single-User gender" depend on " single-user id "column, there is dependent transmission, a sub-sub-table should be out.
4️⃣ fourth normal form: (BC based on the paradigm) multivalued eliminate non-primary property.
Example: the user has a table has "id" and "iphone" and "tel" three columns, id is user 1 are two, respectively,
1 ；188xxxx0000；0796-xx0000
2； 199xxxx0000；0796-00xxxx
The need to establish a deputy table, to eliminate this situation is consistent with the fourth paradigm.
5️⃣ fifth paradigm: (4NF on the basis of) the elimination of ligation-dependent.
Example: a little.
6️⃣ BC Paradigm: (on a 3NF on) all the main properties of the composition of the primary key column of the column in which an arbitrary column entirely dependent on primary keys, eliminate all of the main attributes primary key column of the column in which any row is transmitted to the portion of the main attribute depends.
Examples: "A" and "B" and "C" is the primary key, "A" or "B" or "C" are all primary attributes primary key column of the column in which an arbitrary column , so they must rely entirely on the primary key , but can not exist between the partially dependent.