Interview Question

How can you delete an existing entity without selecting it from the database?

Answer

Entity Framework Core provides several approaches to delete entities without first retrieving them from the database, which can significantly improve performance.

1. Using Remove with a Stub Entity

public async Task DeleteProduct(int productId)
{
    // Create a stub entity with just the primary key
    var product = new Product { Id = productId };
    
    // Attach and remove
    _context.Products.Attach(product);
    _context.Products.Remove(product);
    
    // Save changes
    await _context.SaveChangesAsync();
}

2. Using ExecuteDeleteAsync (EF Core 7.0+)

public async Task DeleteProductWithExecuteDelete(int productId)
{
    // Delete directly without tracking
    await _context.Products
        .Where(p => p.Id == productId)
        .ExecuteDeleteAsync();
}

3. Using Raw SQL for Performance-Critical Deletes

public async Task DeleteProductWithRawSql(int productId)
{
    await _context.Database.ExecuteSqlRawAsync(
        "DELETE FROM Products WHERE Id = {0}", productId);
}

4. Bulk Delete with LINQ and ExecuteDeleteAsync (EF Core 7.0+)

public async Task DeleteExpiredProducts()
{
    // Delete all expired products in one operation
    await _context.Products
        .Where(p => p.ExpiryDate < DateTime.UtcNow)
        .ExecuteDeleteAsync();
}

Key Points 💡

• Deleting without selecting first reduces database roundtrips
• ExecuteDeleteAsync() (EF Core 7+) is the most efficient method
• For older EF Core versions, use Attach() and Remove()
• Raw SQL can be used for complex or bulk deletes
• Consider using transactions for operations that delete multiple related entities
• Be aware of foreign key constraints and cascading deletes
• These methods bypass change tracking and validation logic
• Consider soft deletion for entities that might need to be recovered

Common Follow-up Questions

1. How do these methods handle cascade deletes?
2. What happens if the entity doesn’t exist when using these methods?
3. How would you implement an audit trail for deleted entities?
4. What are the performance implications of bulk deletes on large tables?