Performance Profiling EF Core Queries

Introduction

Performance Profiling EF Core Queries is essential for .NET developers building the data layer of ShopNest.Data — Toolliyo's 100-article EF Core path covering DbContext, Code First, migrations, relationships, LINQ, performance, transactions, and enterprise patterns for SQL Server and cloud deployments.

In Indian delivery projects, teams lose sprints when juniors skip performance profiling ef core queries fundamentals — N+1 queries, missing indexes, sync database calls, or untested migrations. This article prevents that on Performance.

After this article you will

• Explain Performance Profiling EF Core Queries in plain English and in technical EF Core ORM terms
• Implement performance profiling ef core queries in ShopNest.Data (Performance)
• Compare the wrong approach vs the production-ready enterprise approach
• Answer fresher and mid-level EF Core interview questions confidently
• Connect this lesson to Article 88 and the 100-article EF Core roadmap

Prerequisites

• Software: .NET 8 SDK, VS 2022 or VS Code, SQL Server Express / LocalDB
• Knowledge: C# basics
• Previous: Article 86 — Query Logging in EF Core
• Time: 28 min reading + 30–45 min hands-on

Concept deep-dive

Level 1 — Analogy

EF Core is a translator between C# objects and relational tables — you model in C#; it generates SQL and maps rows back to entities.

Level 2 — Technical

Performance Profiling EF Core Queries is part of the EF Core data layer in ShopNest.Data — configure DbContext, keep queries in services/repositories, and test with InMemory or SQLite for Performance.

Level 3 — EF Core data flow

[Application Service / API]
       ▼
[DbContext (Scoped)]
       ▼
[LINQ → Expression Tree → SQL Generator]
       ▼
[SQL Server / PostgreSQL]
       ▼
[Data Reader → Materialization → Change Tracker]
       ▼
[DTO Projection / SaveChangesAsync]

Project structure

ShopNest.Data/
├── ShopNest.Domain/           ← Entity classes
├── ShopNest.Infrastructure/   ← DbContext, configurations, migrations
├── ShopNest.Application/      ← Services, repository interfaces
├── ShopNest.Api/              ← ASP.NET Core host (optional)
└── ShopNest.Tests/            ← Integration tests (SQLite/InMemory)

Hands-on implementation — Performance

Follow the steps below to practice Performance Profiling EF Core Queries in Performance with a minimal working example.

1. Read the lesson objective and list success criteria.
2. Implement the smallest working version.
3. Test happy path and one failure case.
4. Compare your code to the good example below.
5. Note one interview talking point from what you built.

Anti-pattern (quick hack without tests or error handling)

// ❌ BAD — N+1 queries, sync IO, tracked entities returned to API
foreach (var orderId in orderIds)
{
    var order = _context.Orders.Find(orderId); // sync + N round-trips
    dto.Add(Map(order)); // exposes tracked entity graph
}

Production-style example

// ✅ CORRECT — Performance Profiling EF Core Queries on ShopNest (Performance)
public async Task<ProductDto?> GetDtoAsync(int id, CancellationToken ct)
{
    return await _context.Products.AsNoTracking()
        .Where(p => p.Id == id)
        .Select(p => new ProductDto { Id = p.Id, Name = p.Name })
        .FirstOrDefaultAsync(ct);
}

Complete example

await _context.Products
    .Where(p => p.IsPublished)
    .OrderBy(p => p.Name)
    .ToListAsync();

Database design

Product (Id, Name, Price, CategoryId)
Category (Id, Name)
Order (Id, CustomerId, OrderDate, Total)
OrderItem (OrderId, ProductId, Quantity, UnitPrice)

Use FK constraints, indexes on CategoryId and CustomerId, and avoid SELECT * in production LINQ queries.

Query performance

Enable SQL logging in development, add indexes on FK and filter columns, prefer AsNoTracking() for read APIs, and use AsSplitQuery() when Include graphs cause cartesian explosion.

Common errors & fixes

• N+1 queries in loops — Use Include, projection, or explicit loading.
• Tracking large graphs — Use AsNoTracking for read-only queries.
• Ignoring migration reviews — Review generated SQL before applying to production.

Best practices

• 🟢 Register DbContext as Scoped; inject into services, not singletons
• 🟢 Use async LINQ (ToListAsync, SaveChangesAsync) on I/O paths
• 🟡 Use AsNoTracking() for read-only queries and API list endpoints
• 🟡 Review migration SQL before applying to production
• 🔴 Never use EnsureCreated() in shared or production databases
• 🔴 Log generated SQL in dev; monitor slow queries in production

Interview questions

Fresher level

Q1: Explain Performance Profiling EF Core Queries in an EF Core interview.
A: Define the concept, show a ShopNest entity/query example, mention tracking implications, and one production pitfall you avoided.

Q2: Code First vs Database First — when to use which?
A: Code First for greenfield; scaffold from existing DB for legacy; raw SQL/Dapper for hot reporting paths.

Q3: Explain the EF Core query pipeline.
A: LINQ → expression tree → SQL generator → database → data reader → materialization → optional change tracking.

Mid / senior level

Q4: How do you fix N+1 queries?
A: Use Include/projection, split queries, or explicit loading; verify with logged SQL and profiling.

Q5: DbContext lifetime in ASP.NET Core?
A: Register as Scoped — one context per request; never singleton with concurrent requests.

Q6: EF Core vs Dapper vs raw ADO.NET?
A: EF for productivity and change tracking; Dapper/ADO for hand-tuned reads and bulk operations.

Coding round

Write a LINQ query: top 3 customers by total order value on ShopNest orders.

var top = await _context.Orders
    .GroupBy(o => o.CustomerId)
    .Select(g => new { CustomerId = g.Key, Total = g.Sum(o => o.GrandTotal) })
    .OrderByDescending(x => x.Total).Take(3).ToListAsync();

Summary & next steps

• Article 87: Performance Profiling EF Core Queries
• Module: Module 9: Testing & Debugging · Level: ADVANCED
• Applied to ShopNest.Data — Performance

Previous: Query Logging in EF Core
Next: Exception Debugging in EF Core

Practice: Add one small feature using today's pattern — commit with feat(efcore): article-87.

FAQ