Multivalued Dependency (MVD), Fourth Normal Form (4NF), Join Dependency and Fifth Normal Form (5NF) are advanced normalization concepts used when redundancy remains even after BCNF. MVD appears when one attribute independently determines multiple values of another attribute. 4NF removes problematic non-trivial MVDs. Join Dependency explains when a relation can be reconstructed from several projections, and 5NF removes redundancy caused by join dependencies that cannot be solved properly by the lower normal forms. This chapter explains all four ideas from zero with simple examples.
What Comes After BCNF?
Easy order to remember
FD -> BCNF · MVD -> 4NF · Join Dependency -> 5NFIf BCNF is not clear yet, revise BCNF in DBMS: 3NF vs BCNF. Also revise Lossless Join & Dependency Preservation in DBMS because 4NF and 5NF both use decomposition ideas.
What Is Multivalued Dependency (MVD)?
A multivalued dependency exists when one attribute determines a set of values of another attribute, and those values are independent of the remaining attributes. We will write it as X ->> Y.
Suppose one student can have many hobbies and can also know many languages. Hobbies and languages are independent of each other.
Student
Asha
Hobbies + Languages
Hobbies: Chess, Music
Languages: Hindi, English
We can write:
Student ->> Hobby
Student ->> LanguageThe important word is independent. Asha's hobby does not decide which language she knows, and her language does not decide her hobby.
MVD memory line
One X value -> many Y values, independent of the restWhy Multivalued Dependency Creates Repetition
Consider this relation:
| Student | Hobby | Language |
|---|---|---|
| Asha | Chess | Hindi |
| Asha | Chess | English |
| Asha | Music | Hindi |
| Asha | Music | English |
Asha has 2 hobbies and 2 languages. Because the two sets are independent, the table stores all 2 x 2 combinations. If she had 5 hobbies and 4 languages, the table might need 20 rows for one student.
This repetition is not caused by a normal functional dependency. It is caused by two independent multi-valued facts being stored together.
Functional Dependency vs Multivalued Dependency
Functional Dependency
X -> Y
One X value determines one Y value.
Example: Student_ID -> Student_Name.
Multivalued Dependency
X ->> Y
One X value determines a set of Y values that is independent of the remaining attributes.
Example: Student ->> Hobby.
Every functional dependency also implies a corresponding multivalued dependency, but not every multivalued dependency is a functional dependency.
For FD basics, revise Functional Dependency in DBMS.
Trivial and Non-Trivial Multivalued Dependency
X ->> Y is trivial if Y is a subset of X.Example relation:
R(Student, Hobby, Language)Student ->> Hobby is non-trivial because Hobby is not part of Student and Student plus Hobby does not contain Language.
What Is 4NF in DBMS?
A relation is in Fourth Normal Form (4NF) if, for every non-trivial multivalued dependency X ->> Y, X is a superkey.
This is similar to BCNF, but BCNF checks functional dependencies while 4NF also handles multivalued dependencies.
4NF rule
Non-trivial X ->> Y ? Then X must be a superkeyWhy Student-Hobby-Language Violates 4NF
We have:
Student ->> Hobby
Student ->> LanguageStudent is not a superkey of the relation R(Student, Hobby, Language) because one student appears in many rows.
Therefore the non-trivial MVDs violate 4NF.
Before 4NF
STUDENT_INFO(Student, Hobby, Language)
Independent facts are mixed together.
After 4NF
STUDENT_HOBBY(Student, Hobby)
STUDENT_LANGUAGE(Student, Language)
How to Convert a Relation to 4NF
For our example:
STUDENT_HOBBY(Student, Hobby)
STUDENT_LANGUAGE(Student, Language)Now hobby facts and language facts are stored separately, so the cross-product repetition disappears.
Explore MVD, 4NF, Join Dependency and 5NF
MVD: independent sets of values
What to remember
BCNF vs 4NF
| Point | BCNF | 4NF |
|---|---|---|
| Main dependency | Functional Dependency | Multivalued Dependency |
| Rule | Every non-trivial FD determinant is a superkey | Every non-trivial MVD determinant is a superkey |
| Redundancy handled | FD-based redundancy | Independent multi-valued facts |
| Strength | Lower than 4NF | Stronger normal form |
A relation can satisfy BCNF and still violate 4NF because MVD problems can exist even when there is no problematic functional dependency.
What Is Join Dependency in DBMS?
A join dependency says that a relation can be reconstructed exactly by joining several of its projections. It is a more general dependency than multivalued dependency.
Suppose relation R has attribute groups R1, R2 and R3. A join dependency can be written conceptually as:
JD *(R1, R2, R3)It means the original relation R should be exactly equal to the natural join of those projections:
R = R1 JOIN R2 JOIN R3If the join reconstructs R without adding false tuples, the join dependency holds.
Join Dependency memory line
Project into smaller relations -> join them -> get exactly RRelationship Between MVD and Join Dependency
Multivalued dependency can be viewed as a special case of join dependency involving a decomposition into two relation projections.
Join dependency is more general because it can involve three or more projections at the same time.
MVD
Commonly leads to a two-way decomposition.
Handled by 4NF.
Join Dependency
Can require three or more projections.
Handled by 5NF.
What Is 5NF in DBMS?
Fifth Normal Form (5NF), also called Project-Join Normal Form (PJNF), is reached when every non-trivial join dependency is implied by candidate keys.
In simple words, a relation is in 5NF when it cannot be decomposed into smaller relations and then safely reconstructed only because of a non-key join dependency that still creates redundancy.
5NF memory line
No unnecessary redundancy caused by non-key join dependenciesSupplier-Part-Project Example
Consider:
SPJ(Supplier, Part, Project)Suppose a business rule says that valid triples are determined by pairwise relationships:
SP(Supplier, Part)
SJ(Supplier, Project)
PJ(Part, Project)If every valid Supplier-Part-Project triple can be reconstructed exactly by joining these three pairwise relations, then SPJ has a join dependency.
| Relation | Meaning |
|---|---|
| SP | Which supplier can supply which part |
| SJ | Which supplier works on which project |
| PJ | Which part is used in which project |
If the original three-attribute relation stores only facts that can be derived from these three pairwise relations, keeping the big relation can create repeated combinations. Decomposing into SP, SJ and PJ can remove that redundancy.
Why 5NF Problems Are Different From 4NF Problems
4NF problem
Usually two independent sets of multi-valued facts are mixed in one table.
Example: Student-Hobby-Language.
5NF problem
Redundancy appears because a three-way or higher join dependency exists even though lower normal forms may already be satisfied.
Example: Supplier-Part-Project.
When Should You Think About 5NF?
Difference Between 4NF and 5NF
| Point | 4NF | 5NF |
|---|---|---|
| Dependency type | Multivalued Dependency | Join Dependency |
| Main issue | Independent multiple values | Redundancy caused by multi-relation join structure |
| Typical decomposition | Often two relations | Often three or more relations |
| Example | Student-Hobby-Language | Supplier-Part-Project |
| Normal form strength | Stronger than BCNF | Stronger than 4NF |
1NF to 5NF: What Each Normal Form Removes
For the lower forms, revise Normalization in DBMS: 1NF, 2NF & 3NF.
Solved Example: Convert a Relation to 4NF
Given:
R(Student, Hobby, Language)
MVDs:
Student ->> Hobby
Student ->> LanguageStep 1: Check whether the MVDs are non-trivial
Yes. Hobby and Language are separate attributes and neither MVD covers the whole relation with Student.
Step 2: Is Student a superkey?
No. One student can appear in many hobby-language combinations.
Step 3: 4NF verdict
The relation violates 4NF.
Step 4: Decompose
R1(Student, Hobby)
R2(Student, Language)The independent multi-valued facts are now separated.
Solved Example: Why a Supplier-Part-Project Relation May Need 5NF
Suppose:
R(Supplier, Part, Project)and valid triples are completely determined by these pairwise relations:
SP(Supplier, Part)
SJ(Supplier, Project)
PJ(Part, Project)If:
R = SP JOIN SJ JOIN PJand the join produces exactly the valid original triples, then the join dependency holds.
If that non-trivial join dependency is not explained merely by a candidate key, the large relation violates 5NF and can be decomposed into the three smaller pairwise relations.
Lossless Decomposition Still Matters in 4NF and 5NF
Higher normal forms do not mean “split the table as much as possible.” The decomposition should still be lossless.
For 4NF, the standard decomposition using a violating MVD is designed to preserve the original relation through a lossless join.
For 5NF, the join dependency itself describes the projections whose natural join reconstructs the relation.
For the full idea, revise Lossless Join & Dependency Preservation in DBMS.
Mistakes to Avoid
-> and ->> conceptually separate.How to Write MVD, 4NF, JD and 5NF in Semester Exams
Definition answer
Define MVD or 4NF directly and write one notation-based example.
Rule + example
Explain the normal-form rule, show the violating dependency and write the decomposed relations.
Full comparison
Explain MVD, 4NF, join dependency, 5NF, differences, one 4NF decomposition and one Supplier-Part-Project example.
For presentation, revise How to Write High-Scoring Answers in University Semester Exams.
Practice Questions With Answers
1. What is a multivalued dependency?
A multivalued dependency exists when one attribute set determines an independent set of values of another attribute set. It can be written as X ->> Y.
2. What is the difference between FD and MVD?
An FD gives one determined value for a determinant, while an MVD gives an independent set of values.
3. What is 4NF?
A relation is in 4NF if every non-trivial multivalued dependency has a superkey on the left side.
4. Can a relation be in BCNF but not in 4NF?
Yes. BCNF handles functional dependencies, so a relation can satisfy BCNF and still have a problematic multivalued dependency.
5. What is a join dependency?
A join dependency states that a relation can be reconstructed exactly by joining specified projections of that relation.
6. What is 5NF?
5NF, or Project-Join Normal Form, requires every non-trivial join dependency to be implied by candidate keys.
7. What is the difference between 4NF and 5NF?
4NF handles multivalued dependencies, while 5NF handles more general join dependencies.
8. Is every MVD a join dependency?
An MVD can be represented as a special case of join dependency involving a two-way decomposition.
9. Why is 5NF rare?
Most database redundancy is already handled by lower normal forms. 5NF mainly appears in more complex many-to-many-to-many relationship structures.
10. What comes after 5NF?
In most DBMS courses, 5NF is the final commonly taught normal form. Some advanced texts also discuss Domain-Key Normal Form.
One-Minute Revision Sheet
X ->> Y.Ultimate memory line
Independent many-values? Think MVD and 4NF. Complex join rule? Think JD and 5NF.Frequently Asked Questions
What is MVD in DBMS in simple words?
MVD means one attribute determines multiple independent values of another attribute. For example, one student can independently have many hobbies and know many languages.
What is the rule of 4NF?
For every non-trivial multivalued dependency X ->> Y, X must be a superkey.
What is join dependency in DBMS?
Join dependency means a relation can be reconstructed exactly by naturally joining specified projections of that relation.
What is 5NF in DBMS?
5NF, also called Project-Join Normal Form, requires every non-trivial join dependency to be implied by candidate keys.
What is the difference between 4NF and 5NF?
4NF removes redundancy caused by multivalued dependencies, while 5NF handles more general redundancy caused by join dependencies.
Can a table be in BCNF but not in 4NF?
Yes. BCNF checks functional dependencies, while 4NF also checks non-trivial multivalued dependencies.
What Should You Learn Next?
Next: Transactions, ACID Properties & Transaction States
You have now completed the main normalization path from 1NF through 5NF. The next major DBMS block should move to transactions, ACID properties, transaction states, schedules, concurrency control, locking and recovery.
For video revision after reading, use the DBMS YouTube study guide only as a secondary resource.
