Về các phép mã hóa hợp lý trong tiến trình thiết kế mô hình cơ sở dữ liệu

Abstract

Let R=<R⁺,F> be a universal in which  R⁺ is the set of attributes of R, and F is the set of functional dependencies (fds) holding in R. In process of designing of relational scheme, we usually want to pass rapidly, as fast as possible, the procedures as follows:

-To find all key attribute sets of R.

- To find an unredundant cover of F.

- To decompose R into a database (DB) scheme {R_i=<R⁺_i, F_i>,  i=1…n} which satisfies at least 2 conditions:

1/ each f of Fi is an embodied fd,

2/ each Ri satisfies Boyce Code Norm 2, i.e. every nonkey attribute depends functionally on a key of Ri.

At the first stage, the coding stage, one normally assign to each attribute one number, then manipulate on those numbers. Now, we study some method (tricky or logical?) to assign to each “selected” disjoint set a number (a code number). So, we hope not only to reduce the “total attributes” but also to get around with condition of “as long as a century”, even that of “Memory Overflowed”, that is almost indispendable when we are in the process on a microcomputer. For the complexity, on a whole, is a multiple of total attributes, most often.

Also, we present our coding algorithm, the NGH coding algorithm. It not only has low complexity but also produces the least amount of code number-all over logical coding methods.