Customer Management

Overview

Duplicate customer registration with the same email must be prevented. How do you implement email duplicate checking? Loading the entire customer list and comparing manually is inefficient, and code becomes complex as search conditions increase.

This chapter implements Specification pattern to encapsulate search conditions through the Customer domain and validates uniqueness using Repository’s Exists() method. It also covers tracking creation/modification timestamps with the IAuditable interface and the dynamic filter builder pattern.

---

Learning Objectives

After completing this chapter, you will be able to:

1. Encapsulate search conditions with the Specification pattern
2. Leverage Specification composition (And, Or, Not operators)
3. Validate uniqueness with Repository.Exists()
4. Implement a dynamic filter builder (All + conditional chaining)
5. Implement an InMemoryQueryBase-based Query Adapter

---

Core Concepts

Specification Pattern

Encapsulating search conditions as objects makes reuse and composition easy. See how to create individual Specifications and compose them with &, | operators.

// Individual Specification
var emailSpec = new CustomerEmailSpec("kim@example.com");
var nameSpec = new CustomerNameSpec("Kim");

// Composition: & operator for And
var composedSpec = nameSpec & emailSpec;

// Dynamic builder: conditional addition with All as seed
var filter = Specification<Customer>.All;
if (!string.IsNullOrEmpty(nameFilter))
    filter = filter & new CustomerNameSpec(nameFilter);

Specification<T>.All is the identity element for And operations, so when no conditions are added, it returns all data.

Using Specification with Repository

For checking “does data matching this condition exist?” like email duplicate checking, use Exists(). It’s performant because it doesn’t load entire Aggregates.

public interface ICustomerRepository : IRepository<Customer, CustomerId>
{
    FinT<IO, bool> Exists(Specification<Customer> spec);
}

// Email duplicate check
var exists = await repository
    .Exists(new CustomerEmailSpec("kim@example.com"))
    .Run().RunAsync();

---

Project Description

File Structure

Check each file’s role in the Specification pattern.

File	Role
CustomerId.cs	Ulid-based customer identifier
Customer.cs	Customer Aggregate Root (IAuditable)
CustomerDto.cs	Query-side DTO
CustomerEmailSpec.cs	Email Specification (case-insensitive)
CustomerNameSpec.cs	Name Specification (partial match)
ICustomerRepository.cs	Repository + Exists(Specification)
InMemoryCustomerRepository.cs	InMemory Repository implementation
InMemoryCustomerQuery.cs	InMemory Query Adapter

---

Summary at a Glance

A summary of the Specification pattern elements used in this example.

Concept	Implementation
Specification	CustomerEmailSpec, CustomerNameSpec
Composition	spec1 & spec2 (And), spec1 | spec2 (Or)
Identity element	Specification<Customer>.All
Exists validation	ICustomerRepository.Exists(spec)
Query Adapter	InMemoryCustomerQuery : InMemoryQueryBase<Customer, CustomerDto>
Audit tracking	IAuditable -> CreatedAt, UpdatedAt

---

FAQ

Q1: Why not implement Specification as ExpressionSpecification?

A: This example uses an InMemory environment, so only IsSatisfiedBy() is needed. When integrating with EF Core/Dapper, ExpressionSpecification<T> is used to support automatic SQL translation.

Q2: Why use All as the seed in dynamic filters?

A: Specification<T>.All is the identity element for And operations. Since All & X = X, when no conditions are added, all data is returned. This pattern cleanly handles nullable filter parameters.

Q3: Why provide Exists() as a separate method?

A: The intent is clearer than GetById() followed by a null check, and it’s performant because entire Aggregates aren’t loaded. In actual DB environments, it translates to lightweight queries like SELECT COUNT(1).

---

Customer management and the Specification pattern are complete. Next is inventory management. If you delete a product, do the related order histories disappear too? The next chapter explores preserving data while marking it as deleted with the soft delete pattern.

-> Chapter 3: Inventory Management