Raiqub.Expressions.Writing 2.1.7

Expressions

Raiqub.Expressions is a library that provides abstractions for creating specifications and query strategies using LINQ expressions. It also supports querying and writing to databases using various providers.

🏃 Quickstart   |   📗 Guide   |   🔄 Migration

<hr />

Features

• Abstractions for creating specifications and query strategies
• Abstractions for querying and writing to databases
• Supports Entity Framework Core and Marten providers
• Built with .NET Standard 2.0, 2.1, and .NET Core 6.0

NuGet Packages

• Raiqub.Expressions: abstractions for creating specifications
• Raiqub.Expressions.Reading: abstractions for creating query strategies and query sessions and querying from database (defines IDbQuerySessionFactory and IDbQuerySession interfaces)
• Raiqub.Expressions.Writing: abstractions for creating write sessions and writing to database (defines IDbSessionFactory and IDbSession interfaces)
• Raiqub.Expressions.EntityFrameworkCore: implementation of sessions and factories using Entity Framework Core
• Raiqub.Expressions.Marten: implementation of sessions and factories using Marten library

Prerequisites

Before you begin, you'll need the following:

• .NET Standard 2.0 or 2.1, or .NET Core 6.0 installed on your machine
• An IDE such as Visual Studio, Visual Studio Code, or JetBrains Rider
• A database to query against (if using the reading package) or write to (if using the writing package)

Quickstart

To use Raiqub.Expressions in your project, follow these steps:

Entity Framework Core

1. Install the required NuGet package(s) for the database provider you'll be using, such as `Microsoft.EntityFrameworkCore.SqlServer`

2. Install the `Raiqub.Expressions.EntityFrameworkCore` NuGet package

3. Register your DbContext by using `AddDbContextFactory` extension method

   services.AddDbContextFactory<YourDbContext>();
   

4. Register the session and session factories using the appropriate extension method(s) for your database provider:

   services.AddEntityFrameworkExpressions()
       .AddSingleContext<YourDbContext>();
   

Marten

1. Install the `Marten` NuGet package

2. Install the `Raiqub.Expressions.Marten` NuGet package

3. Register the session and session factories using the appropriate extension method(s) for your database provider:

   services.AddMartenExpressions()
       .AddSingleContext();
   

Using

Inject the appropriate session interface (`IDbQuerySession` for read sessions, `IDbSession` for write sessions) into your services, and use it read and write from/to database.

You can also create specifications and query strategies. Here's an example of how to create a simple specification:

public class CustomerIsActive : Specification<Customer>
{
    public override Expression<Func<Customer, bool>> ToExpression()
    {
        return customer => customer.IsActive;
    }
}

And here's an example of how to use the specification:

// session is of type IDbSession or IDbQuerySession and can be injected
var query = session.Query(new CustomerIsActive());
var customers = await query.ToListAsync();

Guide

Creating Specifications

The Specification Pattern is a behavioral design pattern used to encapsulate business rules into composable, reusable and testable objects. This pattern is often used in domains where queries or validation rules need to be expressed in a more readable and maintainable form.

A specification, in the context of this package, is an object that defines a condition that must be satisfied by elements of a certain type. These conditions can be as simple or as complex as needed and are expressed using lambda expressions.

Creating a Simple Specification

The `Raiqub.Expressions` package provides the Specification<T> base class for creating specifications. It is optimized to allow ORM frameworks to evaluate and translate it into SQL queries.

Here's a basic example of a specification that checks if a product is in stock:

public class ProductIsInStock : Specification<Product>
{
    public override Expression<Func<Product, bool>> ToExpression()
    {
        return product => product.AvailableQuantity > 0;
    }
}

Creating Specification Factories

You can also create specification factories, which are static classes that provide predefined specifications. Here's an example of a specification factory for products:

public static class ProductSpecification
{
    public static Specification<Product> IsInStock { get; } =
        Specification.Create<Product>(product => product.AvailableQuantity > 0);

    public static Specification<Product> IsDiscountAvailable(DateTime now) =>
        Specification.Create<Product>(product => product.DiscountStartDate <= now && now <= product.DiscountEndDate);
}

Combining Specifications

Specifications can be combined using extension methods or logical operators to create more complex conditions. This flexibility allows you to express intricate business rules concisely. Here are examples of combining specifications for incidents:

    public static Specification<Incident> IsClosed { get; } =
        Specification.Create<Incident>(incident => incident.Status == IncidentStatus.Closed);

    public static Specification<Incident> IsResolved { get; } =
        Specification.Create<Incident>(incident => incident.Status == IncidentStatus.Resolved);

    // =======================
    // Using extension methods
    // =======================
    public static Specification<Incident> IsNotResolved { get; } =
        IsResolved.Not();

    public static Specification<Incident> IsResolvedOrClosed { get; } =
        IsResolved.Or(IsClosed);

    // =======================
    // Using logical operators
    // =======================
    public static Specification<Incident> IsNotResolved { get; } =
        !IsResolved;

    public static Specification<Incident> IsResolvedOrClosed { get; } =
        IsResolved | IsClosed;

Practical Use Cases

Specifications are valuable for filtering data, composing complex queries, and validating entities against business rules. They are particularly useful when working with Object-Relational Mapping (ORM) frameworks, as they can be translated into SQL queries for efficient database operations.

Creating Query Strategies

The query strategy is based on the Strategy Pattern by defining a strategy for querying the database allowing better concern separation, maintainability and reusability than the repository pattern.

The `Raiqub.Expressions.Reading` package provides abstractions for creating query strategies. You can create a new query strategy by choosing one of several ways available to implement a query strategy.

Single Entity Query

The most common strategy is querying a single entity and for that purpose the interface IEntityQueryStrategy<TSource, TResult> was created and its abstract class implementation EntityQueryStrategy<TSource, TResult>.

Here's an example of a entity query strategy that filters a list of entities based on a set of conditions:

public class GetProductNameQueryStrategy : EntityQueryStrategy<Product, ProductName>
{
    protected override IQueryable<ProductName> ExecuteCore(IQueryable<Product> source)
    {
        return source
            .Where(ProductSpecification.IsInStock)
            .OrderBy(e => e.Name)
            .Select(e => new ProductName { Id = e.Id, Name = e.Name });
    }
}

or, you can define only the preconditions:

public class GetProductInStockQueryStrategy : EntityQueryStrategy<Product>
{
    protected override IEnumerable<Specification<Product>> GetPreconditions()
    {
        yield new ProductIsInStock();
    }
}

or yet, you can create a static class as a provider of query strategies:

public static class ProductQueryStrategy
{
    public static IEntityQueryStrategy<Product, ProductName> GetName() =>
        QueryStrategy.CreateForEntity(
            (IQueryable<Product> source) => source
                .Where(ProductSpecification.IsInStock)
                .OrderBy(e => e.Name)
                .Select(e => new ProductName { Id = e.Id, Name = e.Name });
}

Multiple Entities Query

For the cases where multiple entities need to be queried the interface IQueryStrategy<TResult> was created.

You can implement the interface directly, as the example below:

public class GetProductNameOfOpenStoreQueryStrategy : IQueryStrategy<ProductName>
{
    public IQueryable<TResult> Execute(IQuerySource source) =>
        source => from product in source.GetSetUsing(ProductSpecification.IsInStock)
            join store in source.GetSetUsing(StoreSpecification.IsOpen) on
                product.StoreId equals store.Id
            orderby product.Name
            select new ProductName { Id = e.Id, Name = e.Name };
}

or, can create a static class as a provider of query strategies:

public static class ProductQueryStrategy
{
    public static IQueryStrategy<ProductName> GetNameOfOpenStore() =>
        QueryStrategy.Create(
            source => from product in source.GetSetUsing(ProductSpecification.IsInStock)
                join store in source.GetSetUsing(StoreSpecification.IsOpen) on
                    product.StoreId equals store.Id
                orderby product.Name
                select new ProductName { Id = e.Id, Name = e.Name });
}

Creating Query Sessions and Querying Data

To create a query session and query data using a query strategy, follow these steps:

1. Inject an instance of IDbQuerySessionFactory into your service or controller.
2. Use the Create() method of the IDbQuerySessionFactory interface to create a new query session.
3. Call the Query() method on the query session, passing in your query strategy or specification instance.
4. Call one of the methods on the resulting IDbQuery<T> interface to execute the query and retrieve the results.

await using (var session = querySessionFactory.Create())
{
    IDbQuery<Customer> query = session.Query(new CustomerIsActive());
    IReadOnlyList<Customer> customers = await query.ToListAsync();
}

Creating Write Sessions and Writing Data

To create a write session and write data to the database, follow these steps:

1. Inject an instance of IDbSessionFactory into your service or controller.
2. Use the Create() method of the IDbSessionFactory interface to create a new write session.
3. Call the appropriate methods on the write session to perform insert, update, or delete operations on your entities.
4. Call the SaveChangesAsync() method on the write session to persist your changes to the database.

await using (var session = sessionFactory.Create())
{
    var blog = new Blog { Url = "https://example.com" };
    session.Add(blog);
    await session.SaveChangesAsync();
}

Defining custom SQL query for entity (Entity Framework)

Sometimes, you may need to define custom SQL queries to retrieve entities from the database in Entity Framework. This can be useful when you have complex queries that cannot be easily expressed using the standard LINQ queries. Here are the steps to define a custom SQL query for an entity:

Step 1: Create a Custom SQL Provider

To define a custom SQL query, start by creating a class that implements the ISqlProvider<TEntity> interface. This interface defines a method, GetQuerySql(), where you can specify your custom SQL query.

private class BlogSqlProvider : ISqlProvider<Blog>
{
    public SqlString GetQuerySql() => SqlString.FromSqlInterpolated($"SELECT \"Id\", \"Name\" FROM \"Blog\"");
}

In this example, we've created a BlogSqlProvider class that implements the ISqlProvider<Blog> interface and defines a custom SQL query for retrieving blog entities.

Step 2: Implement the GetQuerySql() Method

Inside your custom SQL provider class, implement the GetQuerySql() method. This method should return a SqlString instance containing your raw or interpolated SQL query. Customize the query to match your specific requirements and entity structure.

Step 3: Register the Custom SQL Provider

To make your custom SQL provider available for dependency injection, you need to register it with your application's service container. You need to register it using the singleton lifetime, as show below:

services.AddSingleton<ISqlProvider, BlogSqlProvider>();

In this registration, we specify the interface ISqlProvider and its corresponding implementation BlogSqlProvider. This allows the Raiqub.Expressions to resolve and use your custom SQL provider when needed.

Best Practices

• Keep Queries Simple: While custom SQL queries provide flexibility, it's essential to keep them as simple and readable as possible. Avoid complex queries that can be hard to maintain.
• Use Parameters: If your custom query requires input parameters, use parameterized queries to prevent SQL injection and improve performance.
• Test Thoroughly: Custom SQL queries may bypass some of Entity Framework's built-in protections. Ensure thorough testing to validate the correctness and security of your queries.

Split queries (Entity Framework)

When working with Entity Framework, you may encounter situations where loading related entities in a single query can lead to a "cartesian explosion". This happens when multiple related entities are included in the query, and the resulting dataset becomes excessively large due to cartesian products, impacting performance.

To mitigate this issue, Entity Framework Core allows you to specify the use of split queries for loading related entities. Split queries separate the loading of each related entity into its own query, preventing the cartesian explosion problem.

You can enable split queries for a specific entity by configuring the query options in your application's service configuration. Here's how you can do it:

services.AddEntityFrameworkExpressions()
    .Configure<Blog>(options => options.UseSplitQuery = true)
    .AddSingleContext<YourDbContext>();

In this example:

• services.AddEntityFrameworkExpressions() registers the necessary services for Entity Framework Expressions.
• .Configure<Blog>(options => options.UseSplitQuery = true) configures the split query option for the Blog entity. Replace Blog with the entity type for which you want to enable split queries.
• .AddSingleContext<YourDbContext>() adds the Entity Framework context to your application.

When to Use Split Queries

Split queries are particularly useful when you have large datasets or complex relationships between entities, and you want to optimize performance by avoiding the cartesian explosion issue. However, enabling split queries should be considered on a case-by-case basis, as it may introduce additional database queries.

Keep in mind that using split queries is just one of many performance optimization techniques available in Entity Framework. Careful consideration of your data model, query patterns, and database design is essential for achieving optimal performance.

By configuring split queries when appropriate, you can strike a balance between efficient data retrieval and preventing performance bottlenecks caused by cartesian explosions.

For more information refer to Single vs. Split Queries article.

Supported Databases

Currently, Raiqub.Expressions supports the following ORM libraries:

• Entity Framework Core
• Marten

If you need to use another ORM library, you will need to implement your own database session factory and database session implementing `IDbSessionFactory` and `IDbSession` interfaces.

Migration Guide

Key Changes in 2.0.0

In the Version 2.0.0 release of our library, the IQueryModel-related interfaces and classes was renamed to IQueryStrategy. These changes are summarized in the table below:

We chose to rename these interfaces and classes to use the "Strategy" suffix (IQueryStrategy, IEntityQueryStrategy, etc.) to better reflect their purpose. This new naming convention aligns with design patterns like the Strategy Pattern, which emphasizes encapsulating algorithms and making them interchangeable.

It's important to note that the IEntityQueryStrategy interface now extends the IQueryStrategy interface. This hierarchy reflects the relationship between entity-specific query strategies and general query strategies. You can use this inheritance structure to create custom query strategies that build upon the foundation provided by IQueryStrategy.

Contributing

If something is not working for you or if you think that the source file should change, feel free to create an issue or Pull Request. I will be happy to discuss and potentially integrate your ideas!

License

This library is licensed under the MIT License.