N O R M A L IZ AT IO N
O F RE L AT IO N A L DATA B A S E S
MIT 202 – Adv a n c e d D ata b a s e M a n a g e m e n t Sys te m
Mr. Jos e p h S. Joa q u in
O b je ctive s
At the end of this top ic , stu d e n ts will be able to:
• D e te rm in e the ste p s on how to con v e rt use r ’s vie w
of info rm a tio n to a rela tio n a l data b a s e ; and
• C re a te a rela tio n a l data b a s e usin g the prin c ip le s of
norm a liz a tio n .
G u id e Qu estion s
• H o w to cre a te a rela tio n a l data b a s e fro m use r ’s
vie w of info rm a tio n ?
W h a t is Norm a liza tio n ?
N o rm a liz a tio n is the pro c e s s of eva lu a tin g and
corre c tin g tab le stru c tu re s to m in im iz e data
red u n d a n c ie s and to elim in a te ano m a lie s in add in g ,
upd a tin g , and dele tio n of rec o rd s .
No rm al Fo rm s
• 1N F (Firs t N o rm a l Form )
• 2N F (S e c o n d N o rm a l Fo rm )
• 3N F (Th ird N o rm a l Form )
• B C N F (B o y c e -C o d d N o rm a l Form )
Le ve ls of No rm al Fo rm s
2N F is bette r tha n 1N F and 3N F is m u c h bette r tha n
1N F and 2N F.
M o s t bus in e s s -re la te d data b a s e s use d as the hig h e s t
lev e l in norm a liz a tio n pro c e s s .
Exam p le of User ’s View of
Info rm a tio n
N o te : E xa m in e the follo w in g exa m p le as you r
refe re n c e in norm a liz a tio n pro c e d u re .
First No rm al Fo rm
To con v e rt use r ’s vie w of info rm a tio n to 1N F,
rep e a tin g gro u p s m u s t be elim in a te d w h ile deriv in g
its nam e from gro u p of m u ltip le entrie s tha t m a y exis t
in any sin g le key attrib u te .
R e la tio n a l tab le m u s t not con ta in rep e a tin g gro u p s
and norm a liz in g the tab le stru c tu re will red u c e data
red u n d a n c ie s .
First No rm al Fo rm (Continued)
Thre e -s te p pro c e d u re :
• Ste p 1 – Elim in a te R e p e a tin g Gro u p s
• Ste p 2 – Ide n tify the Prim a ry K e y s
• Ste p 3 – Ide n tify All D e p e n d e n c ie s
First No rm al Fo rm (Continued)
For Ste p 1, pre s e n t data in a tab u la r form a t, w h e re
eac h cell has a sin g le valu e and the re are no
rep e a tin g gro u p s .
Elim in a te rep e a tin g gro u p s by rem o v in g nulls to
m a k e sure tha t eac h rep e a tin g gro u p attrib u te
con ta in s an app ro p ria te data valu e .
First No rm al Fo rm (Continued)
E xa m p le
First No rm al Fo rm (Continued)
For Ste p 2, ide n tify the prim a ry key s tha t uniq u e ly
ide n tify an attrib u te valu e .
For Ste p 3, ide n tify the pos s ib le fun c tio n a l
dep e n d e n c ie s thro u g h a dep e n d e n c y dia g ra m .
First No rm al Fo rm (Continued)
E xa m p le
The arro w s abo v e the attrib u te s indic a te des ira b le
dep e n d e n c ie s , tho s e that are base d on the prim a ry key and the
arro w s belo w the attrib u te s indica te less des ira b le
dep e n d e n c ie s .
First No rm al Fo rm (Continued)
P a rtia l dep e n d e n c ie s are dep e n d e n t on prim a ry key
w h ile tra n s itiv e dep e n d e n c ie s are dep e n d e n c ie s of
one non -p rim e attrib u te on ano th e r non -p rim e
attrib u te tha t can pro d u c e data ano m a lie s .
Fun c tio n a l D e p e n d e n c y
PRO J _ N U M + EM P _ N U M PRO J _ N A M E , EM P _ N A M E , JO B _ C L A S S , CHG _ H O U R S ,
HO U R S
P a rtia l D e p e n d e n c y
PR O J _ N U M PR O J _ N A M E
EM P _ N U M EM P _ N A M E , JO B _ C L A S S , CH G _ H O U R
First No rm al Fo rm (Continued)
Tra n s itiv e D e p e n d e n c y
JO B _ C LA S S C H G _H O U R
First No rm al Fo rm (Continued)
The tab le is in 1N F if:
• All key attrib u te s has bee n defin e d ;
• The re is no rep e a tin g gro u p s in the tab le ; and
• All attrib u te s are de p e n d e n t on prim a ry key
A n d the 1N F dep e n d e n c y dia g ra m sho w s tha t all
rela tio n a l tab le s satis fy the 1N F req u ire m e n ts .
First No rm al Fo rm (Continued)
Ta b le s con ta in partia l de p e n d e n c ie s tha t are :
• D e p e n d e n c ie s bas e d on only part of the prim a ry
key ;
• D e p e n d e n c ie s use d for perfo rm a n c e rea s o n s bu t
with cau tio n s ; and
• D e p e n d e n c ie s w h ic h are sub je c t to red u n d a n c ie s .
Secon d No rm a l Form
The des ig n of rela tio n a l data b a s e can be im p ro v e d
by con v e rtin g the 1N F to 2N F.
Ste p s to con v e rt 1N F to 2N F :
• Ste p 1. Ide n tify All K e y C o m p o n e n ts
• Ste p 2. Ide n tify the D ep e n d e n t Attrib u te s
Secon d No rm a l Form (Continued)
For Ste p 1, to ide n tify the key com p o n e n ts , write
eac h key on a sep a ra te line and write the orig in a l
(co m p o s ite ) key on las t line .
E xa m p le :
• PRO J_N U M
• EMP_N U M
• PRO J_N U M , EMP _N U M
Note : This key com p o n e n t will be used on its corre sp o n d in g table .
Secon d No rm a l Form (Continued)
For Ste p 2, to ide n tify the dep e n d e n t attrib u te s , from
1N F D e p e n d e n c y Dia g ra m , dete rm in e w hic h
attrib u te s are dep e n d e n t on oth e r attrib u te s .
E xa m p le
• P R O J E C T (P R O J _ N U M , P R O J _ N A M E )
• E M P L O Y E E (E M P _ N U M , E M P _ N A M E , JO B _ C L A S S ,
CHG _H O U R S )
• AS S IG N (P R O J _ N U M , E M P _ N U M , A S S IG N _ H O U R S )
Secon d No rm a l Form (Continued)
2N F D e p e n d e n c y Dia g ra m
Secon d No rm a l Form (Continued)
The tab le is in 2N F if:
• It is in 1N F ; and
• The re is no partia l dep e n d e n c y and no attrib u te is
dep e n d e n t on only a po rtio n of the prim a ry key.
Th ird No rm al Form
D a ta ano m a lie s in any raw sou rc e are elim in a te d by
con v e rtin g 2N F des ig n to 3N F.
Ste p s to con v e rt 2N F to 3N F :
• Ste p 1. Ide n tify E ac h N e w D e te rm in a n t;
• Ste p 2. Ide n tify the D ep e n d e n t Attrib u te s ; and
• Ste p 3. R e m o v e the dep e n d e n t attrib u te s from
tran s itiv e dep e n d e n c ie s .
Th ird No rm al Form (Continued)
For Ste p 1, to ide n tify a new dete rm in a n t, w rite the
dete rm in a n t of tran s itiv e dep e n d e n c y as the prim a ry
key of the new tab le (de te rm in a n t is an attrib u te
w h o s e valu e de te rm in e s oth e r valu e s with in a row ).
For Ste p 2, ide n tify the dep e n d e n t attrib u te s of
dete rm in a n t ide n tifie d in Ste p 1.
E xa m p le
• JO B _ C L A S S → C H G _ H O U R
Th ird No rm al Form (Continued)
A n d giv e a nam e for the new tab le tha t refle c ts its
con te n t and fun c tio n .
E xa m p le
• JO B (JO B _ C L A S S , C H G _ H O U R )
Th ird No rm al Form (Continued)
For Ste p 3, to rem o v e the dep e n d e n t attrib u te s from
tran s itiv e dep e n d e n c ie s :
• Elim in a te all dep e n d e n t attrib u te s in tran s itiv e
rela tio n s h ip (s ) from eac h tab le ;
• D ra w a new de p e n d e n c y dia g ra m to sho w all tab le s
defin e d in in Ste p 1 to 3; and
• C h e c k the new and m o d ifie d tab le s in Ste p 3 to
m a k e sure tha t eac h has a dete rm in a n t and doe s
not con ta in ina p p ro p ria te de p e n d e n c ie s .
Th ird No rm al Form (Continued)
B y rem o v in g the dep e n d e n t attrib u te s the res u lt will
hav e the follo w in g tab le s :
• P R O J E C T (P R O J _ N U M , P R O J _ N A M E )
• E M P L O Y E E (E M P _ N U M , E M P _ N A M E ,
JO B _ C L A S S )
• A S S IG N (P R O J _ N U M , E M P _ N U M ,
A S S IG N _ H O U R S )
• JO B (JO B _ C L A S S , C H G _ H O U R S )
Th ird No rm al Form (Continued)
3N F D e p e n d e n c y Dia g ra m
Th ird No rm al Form (Continued)
To che c k if the data b a s e is in 3N F, if the des ig n is in
2N F (w ith no partia l dep e n d e n c ie s ) and has no
tran s itiv e dep e n d e n c ie s .
Im p rovem e n t in Da tab a se De sig n
Ta b le stru c tu re s are cle a n e d to elim in a te the
trou b le s o m e initia l, partia l, and tran s itiv e
dep e n d e n c ie s . N o rm a liz a tio n is valu a b le bec a u s e its
prin c ip le to elim in a te data red u n d a n c ie s .
H o w e v e r, to m a k e a goo d data b a s e des ig n , nob o d y
can rely on norm a liz a tio n pro c e s s but from the
initia tiv e and und e rs ta n d in g of a pers o n w o rk in g
beh in d the des ig n .
End
