150 lines
6.7 KiB
Markdown
150 lines
6.7 KiB
Markdown
## **Choosing the right datatype for integer values, specially IDs (Chat GPT):**
|
||
|
||
### **1️⃣ Integer Data Types (int, short, long) in C# and SQL Server**
|
||
|
||
| **C# Type** | **SQL Server Type** | **Size** | **Range** |
|
||
| ----------- | ------------------- | -------- | ------------------------------------------------------- |
|
||
| `byte` | `TINYINT` | 1 byte | 0 to 255 |
|
||
| `short` | `SMALLINT` | 2 bytes | -32,768 to 32,767 |
|
||
| `int` | `INT` | 4 bytes | -2,147,483,648 to 2,147,483,647 |
|
||
| `long` | `BIGINT` | 8 bytes | -9,223,372,036,854,775,808 to 9,223,372,036,854,775,807 |
|
||
|
||
❗ **Important Notes:**
|
||
|
||
- The **ranges are the same in C# and SQL Server** because both use the same underlying storage.
|
||
- `SMALLINT` and `TINYINT` **save space**, but be careful about hitting the limit.
|
||
- `BIGINT` is needed only if you expect **billions** of records.
|
||
---
|
||
|
||
### **2️⃣ What to Use for User ID, Ticket ID, Gender ID? (Some examples)**
|
||
|
||
| **Field** | **Recommended C# Type** | **SQL Server Type** | **Why?** |
|
||
| ----------- | ----------------------- | ----------------------- | ----------------------------------------------------------- |
|
||
| `UserId` | `int` or `long` | `INT` or `BIGINT` | `INT` is usually enough unless expecting billions of users. |
|
||
| `TicketId` | `int` or `long` | `INT` or `BIGINT` | Use `BIGINT` if expecting massive ticket volumes. |
|
||
| `GenderId` | `byte` or `short` | `TINYINT` or `SMALLINT` | Gender options are limited, so `TINYINT` is sufficient. |
|
||
| `CompanyId` | `int` | `INT` | Companies are limited, `INT` is fine. |
|
||
| `VehicleId` | `int` | `INT` | Use `INT`, as vehicle count is manageable. |
|
||
| `Price` | `decimal(18,2)` | `DECIMAL(18,2)` | Avoid `float`/`double` due to rounding issues. |
|
||
|
||
---
|
||
|
||
### **3️⃣ Should I Use GUIDs for User IDs or Ticket IDs?**
|
||
|
||
- **Use `GUID` (`UNIQUEIDENTIFIER`) for IDs only if:**
|
||
- Data is distributed across multiple databases.
|
||
- Security is critical (e.g., preventing sequential guessing of IDs).
|
||
- Otherwise, **stick with `int` or `long`** for performance.
|
||
|
||
📌 **Example in C# (EF Core Model):**
|
||
|
||
```csharp
|
||
public class Ticket
|
||
{
|
||
public int TicketId { get; set; } // Primary key
|
||
public int UserId { get; set; } // Foreign key
|
||
public decimal Price { get; set; } // Use decimal for money
|
||
public DateTime PurchaseDate { get; set; }
|
||
}
|
||
```
|
||
|
||
📌 **Fluent API (SQL Mapping)**
|
||
|
||
```csharp
|
||
protected override void OnModelCreating(ModelBuilder modelBuilder)
|
||
{
|
||
modelBuilder.Entity<Ticket>()
|
||
.Property(t => t.Price)
|
||
.HasColumnType("DECIMAL(18,2)");
|
||
}
|
||
```
|
||
|
||
|
||
|
||
## **`required` keyword (Chat GPT) :
|
||
|
||
C# **IntelliSense** suggests adding `required` to string properties because of **nullable reference types (NRT)** introduced in **C# 8.0+**.
|
||
|
||
### **1️⃣ What Does `required` Do?**
|
||
|
||
- `required` **forces initialization** of the property when creating an object.
|
||
- It is **not a data annotation** (like `[Required]` in EF Core), but a **C# keyword** that affects **compile-time checks**.
|
||
|
||
📌 **Example Without `required`**
|
||
|
||
```csharp
|
||
public class User
|
||
{
|
||
public string Name { get; set; } // Warning: "Non-nullable property 'Name' is uninitialized"
|
||
}
|
||
```
|
||
|
||
🔴 **Problem**: The compiler warns that `Name` is not initialized.
|
||
✅ **Fix**: Add `required` or initialize the property.
|
||
|
||
📌 **Example With `required`**
|
||
|
||
```csharp
|
||
public class User
|
||
{
|
||
public required string Name { get; set; } // No warning
|
||
}
|
||
```
|
||
|
||
✅ **Effect**: You **must** provide `Name` when creating a `User` object.
|
||
|
||
```csharp
|
||
var user = new User { Name = "Mehrdad" }; // ✅ Works
|
||
var invalidUser = new User(); // ❌ Compilation Error: Name is required
|
||
```
|
||
|
||
|
||
|
||
## **A note about strings in C# (Chat GPT) :**
|
||
**`string` is nullable in C#**, but in **nullable reference types (C# 8+), `string` is treated as non-nullable unless explicitly marked `string?`**.
|
||
- **`string`** → Default behavior (non-nullable by default in nullable context).
|
||
- **`string?`** → Explicitly nullable.
|
||
|
||
## **Why virtual navigation properties? (Chat GPT):**
|
||
|
||
- If you mark a navigation property as `virtual`, EF Core **creates a proxy class** at runtime that overrides the property and loads related data **only when accessed**.
|
||
- This is called **Lazy Loading**, meaning data is not fetched until needed.
|
||
- If you don’t mark it as `virtual`, you **must** load relationships using `.Include()` (Eager Loading).
|
||
|
||
### **What Type Should Navigation Properties Be?**
|
||
|
||
| **Scenario** | **Recommended Type** | **Why?** |
|
||
| --------------------------------------------------------------------- | ----------------------------- | --------------------------------------------------------------------- |
|
||
| **Single reference** (e.g., `Ticket → Transportation`) | `virtual Transportation` | Represents a **one-to-one** or **many-to-one** relationship. |
|
||
| **Collection of related entities** (e.g., `Transportation → Tickets`) | `virtual ICollection<Ticket>` | Best for **one-to-many** relationships, supports lazy loading. |
|
||
| **Alternative for collections** | `virtual List<Ticket>` | Works the same, but **EF prefers `ICollection<T>`**. |
|
||
| **Using `IEnumerable<T>`** | ❌ **Avoid** | EF **does not recognize** `IEnumerable<T>` for navigation properties. |
|
||
### **One Scenario to look after if using lazy loading:**
|
||
#### **What is the N+1 Query Problem?**
|
||
|
||
The **N+1 query problem** happens when EF Core **makes too many separate database queries** instead of loading data efficiently.
|
||
|
||
##### **Example Scenario**
|
||
|
||
Let’s say you have **100 tickets**, and each ticket has a related **Transportation** entity.
|
||
|
||
You run this code:
|
||
|
||
```c#
|
||
var tickets = context.Tickets.ToList(); // Loads all tickets foreach (var ticket in tickets)
|
||
{
|
||
Console.WriteLine(ticket.Transportation.Name); // Lazy loads Transportation for each ticket
|
||
}
|
||
```
|
||
|
||
##### **What Happens?**
|
||
|
||
1. **1 Query:** EF Core first loads all `Tickets`.
|
||
2. **N Queries:** Then, for each **Ticket**, EF Core makes a separate query to fetch `Transportation` (so 100 additional queries).
|
||
3. **Total Queries:** **1 + 100 = 101 queries!** 🚨 **Bad performance!**
|
||
|
||
---
|
||
|
||
|
||
|