Validation and Transformation

What is Validation?

Validation is the step where incoming data is checked before business logic runs. It ensures the request is safe, correct, and meaningful.

In a backend flow, validation usually happens after the request hits the API and before it reaches controllers/services.

Why Validation is Important

• Prevents bad or malicious data
• Avoids runtime errors
• Keeps business logic clean
• Improves API reliability

Types of Validation

1. Syntactic Validation (Structure Check)

Checks how data looks.

Examples:

• Required fields are present
• Correct JSON format
• String length, min/max values

{
  "email": "user@example.com",
  "age": 25
}

✔ email exists ✔ age is a number

❌ Missing fields or invalid format → rejected here

2. Type Validation (Data Type Check)

Ensures values are of the expected type.

Examples:

• age must be a number
• isAdmin must be boolean
• createdAt must be a date

age: number
isActive: boolean

❌ age: "twenty" → invalid

3. Semantic Validation (Meaning Check)

Checks if data makes sense in real life.

Examples:

• Email must be unique
• User must exist in DB
• Start date must be before end date

startDate < endDate

✔ Structure is valid ✔ Types are valid ❌ Logic is invalid → rejected here

What is Transformation?

Transformation means changing data into a format your system understands and uses internally.

It happens after validation and before business logic.

Why Transformation is Needed

• Convert raw input into domain models
• Normalize data
• Remove unnecessary fields
• Prepare data for services & database

Common Transformations

1. Type Conversion

"25" → 25
"true" → true

2. Rename Fields

user_name → userName

3. Add Derived Fields

fullName = firstName + lastName

4. Remove Extra Fields

{
  ;(email, password)
} // drop unknown fields

5. Format Standardization

email → lowercase
phone → E.164 format

Validation + Transformation Flow (Mermaid)

Where This Fits in Your Diagram

• Validation Layer

  • Syntactic
  • Type
  • Semantic

• Transformation Layer

  • DTO → Domain Model
  • Clean & normalized data

• Controller

  • No validation logic
  • Just orchestration

• Service

  • Pure business rules

Simple Rule to Remember

Validate for correctness, transform for usefulness.