PM> Install-Package Shuttle.Core.Data
Provides an abstraction built directly on ADO.NET which falls within the Micro ORM space.
NOTE: Since a database connection is represented by a IDatabaseContext instance it is important to understand that this instance is not thread-safe. It is therefore important to ensure that the IDatabaseContext instance is not shared between threads. See the DatabaseContextScope to ensure thread-safe database context flow.
The Shuttle.Core.Data package provides a thin abstraction over ADO.NET by making use of the DbProviderFactories. Even though it provides object/relational mapping mechanisms it is in no way a fully fledged ORM.
Connections may be added by providing all the required information:
services.AddDataAccess(builder =>
{
builder.AddConnection(name, providerName, connectionString);
});A connection may also be added by omitting the connectionString, in which case it will be read from the ConnectionStrings section:
services.AddDataAccess(builder =>
{
builder.AddConnectionString(name, providerName);
});The relevant options may be set using the builder:
services.AddDataAccess(builder =>
{
builder.Options.CommandTimeout = timeout;
builder.Options.DatabaseContextFactory.DefaultConnectionStringName = "connection-string-name";
});The default JSON settings structure is as follows:
{
"Shuttle": {
"DataAccess": {
"CommandTimeout": 25,
"DatabaseContextFactory":
{
"DefaultConnectionStringName": "connection-string-name",
}
}
}
}The DatabaseContextService contains a collection of the DatabaseContext instances created by IDatabaseContextFactory. However, since the DatabaseContextService is a singleton the same collection will be used in all thread contexts. This includes not only the same execution context, but also "peered" execution context running in parallel.
To enable an individual execution/thread context-specific collection which also enables async context flow wrap the initial database context creation in a new DatabaseContextScope():
using (new DatabaseContextScope())
{
// database interaction
})In order to access a database we need a database connection. A database connection is represented by an IDatabaseContext instance that may be obtained by using an instance of an IDatabaseContextFactory implementation.
The DatabaseContextFactory implementation makes use of an IDbConnectionFactory implementation which creates a System.Data.IDbConnection by using the provider name and connection string, which is obtained from the registered connection name. An IDbCommandFactory creates a System.Data.IDbCommand by using an IDbConnection instance.
var databaseContextFactory = provider.GetRequiredService<IDatabaseContextFactory>();
using (var databaseContext = databaseContextFactory.Create("connection-name"))
{
// database interaction
}
// or, in async/await implementations
using (new DatabaseContextScope())
using (var databaseContext = databaseContextFactory.Create("connection-name"))
{
// database interaction that will flow across threads
}An IQuery encapsulates a database query that can be executed:
void Prepare(IDbCommand command);This should ensure that the given IDbCommand is configured for execution by setting the relvant command attributes and parameters.
IQuery AddParameter(IColumn column, object value);This method is used to add a parameter to the query. The IColumn instance is used to define the column type and the value is the value that should be used for the parameter.
The Query represents a Text command type:
public Query(string commandText, CommandType commandType = CommandType.Text)You can then add parameters to the query:
query.AddParameter(new Column<Guid>("Id", DbType.Guid), new Guid('{75208260-CF93-454E-95EC-FE1903F3664E}'));Typically you would not want to create a Column each time you need it and these are also quite fixed. A column mapping can, therefore, by defined statically:
using System;
using System.Data;
using Shuttle.Core.Data;
namespace Shuttle.Ordering.DataAccess
{
public class OrderColumns
{
public static readonly Column<Guid> Id =
new Column<Guid>("Id", DbType.Guid);
public static readonly Column<string> OrderNumber =
new Column<string>("OrderNumber", DbType.String, 20);
public static readonly Column<string> OrderDate =
new Column<string>("OrderDate", DbType.DateTime);
public static readonly Column<string> CustomerName =
new Column<string>("CustomerName", DbType.String, 65);
public static readonly Column<string> CustomerEMail =
new Column<string>("CustomerEMail", DbType.String); // size omitted
}
}There are quite a few options that you can set on the Column in order to represent your column properly.
public T Value(DataRow row)This will return the typed value of the specified column as contained in the passed-in DataRow.
The DatabaseGateway is used to execute IQuery instances in order return data from, or make changes to, the underlying data store. If there is no active open IDatabaseContext returned by the DatabaseContextService.Current an InvalidOperationException will be thrown.
The following sections each describe the methods available in the IDatabaseGateway interface.
Task<IDataReader> GetReaderAsync(IQuery query, CancellationToken cancellationToken = default);Returns an IDataReader instance for the given query statement:
using (databaseContextFactory.Create("connection-name"))
{
var reader = await gateway.GetReaderAsync(new Query("select Id, Username from dbo.Member"));
}Task<int> ExecuteAsync(IQuery query, CancellationToken cancellationToken = default);Executes the given query and returns the number of rows affected:
using (databaseContextFactory.Create("connection-name"))
{
await gateway.ExecuteAsync(new Query("delete from dbo.Member where Username = 'mr.resistor'"));
}Task<T> GetScalarAsync<T>(IQuery query, CancellationToken cancellationToken = default);Get the scalar value returned by the select query. The query shoud return only one value (scalar):
using (var databaseContext = databaseContextFactory.Create("connection-name"))
{
var username = await gateway.GetScalarAsync<string>(new Query("select Username from dbo.Member where Id = 10"));
var id = await gateway.GetScalarAsync<int>(new Query.Create("select Id from dbo.Member where Username = 'mr.resistor'") );
}Task<DataTable> GetDataTableAsync(IQuery query, CancellationToken cancellationToken = default);Returns a DataTable containing the rows returned for the given select query.
using (databaseContextFactory.Create("connection-name"))
{
var table = await gateway.GetDataTableAsync(new Query("select Id, Username from dbo.Member"));
}Task<IEnumerable<DataRow>> GetRowsAsync(IQuery query, CancellationToken cancellationToken = default);Returns an enumerable containing the DataRow instances returned for a select query:
using (databaseContextFactory.Create("connection-name"))
{
var rows = await gateway.GetRowsAsync(new Query("select Id, Username from dbo.Member"));
}Task<DataRow> GetRowAsync(IQuery query, CancellationToken cancellationToken = default);Returns a single DataRow containing the values returned for a select statement that returns exactly one row:
using (databaseContextFactory.Create("connection-name"))
{
var row = await gateway.GetRowAsync(new Query("select Id, Username, EMail, DateActivated from dbo.Member where Id = 10") );
}An IDataRepository<T> implementation is responsible for returning a hydrated object. To this end you make use of the DataReposity<T> class that takes a IDatabaseGateway instance along with a IDataRowMapper<T> used to create the hydrated instance.
The following methods can be used to interact with your object type.
Task<IEnumerable<T>> FetchItemsAsync(IQuery query, CancellationToken cancellationToken = default);Uses the select clause represented by the IQuery instance to create a list of objects of type T. The select clause will need to select all the required columns and will, typically, return more than one instance.
Task<T> FetchItemAsync(IQuery query, CancellationToken cancellationToken = default);Returns a single object instance of type T that is hydrated using the data returned from the select clause represented by the IQuery instance.
Task<MappedRow<T>> FetchMappedRowsAsync(IQuery query, CancellationToken cancellationToken = default);This is similar to the FetchItems method but instead returns a list of MappedRow<T> instances. Uses the select clause represented by the IQuery instance to create a list of MappedRow instances of type T. The select clause will need to select all the required columns and will, typically, return more than one instance.
Task<IEnumerable<MappedRow<T>>> FetchMappedRowAsync(IQuery query, CancellationToken cancellationToken = default);Similar to the FetchItem method but instead return a MappedRow<T> instance that is hydrated using the data returned from the select clause represented by the IQuery instance.
Task<bool> ContainsAsync(IQuery query, CancellationToken cancellationToken = default);Returns true is the IQuery instance select clause returns an int scalar that equals 1; else returns false.
The Query enables you to create any query using the native language structure:
var query = new Query("select UserName from dbo.Member where Id = @Id")
.AddParameter(new Column<Guid>("Id", DbType.Guid), new Guid('{75208260-CF93-454E-95EC-FE1903F3664E}'));You use this interface to implement a mapper for a DataRow that will result in an object of type T:
using System.Data;
using Shuttle.Core.Data;
using Shuttle.Process.Custom.Server.Domain;
namespace Shuttle.ProcessManagement
{
public class OrderProcessMapper : IDataRowMapper<OrderProcess>
{
public MappedRow<OrderProcess> Map(DataRow row)
{
var result = new OrderProcess(OrderProcessColumns.Id.Value(row))
{
CustomerName = OrderProcessColumns.CustomerName.Value(row),
CustomerEMail = OrderProcessColumns.CustomerEMail.Value(row),
OrderId = OrderProcessColumns.OrderId.Value(row),
InvoiceId = OrderProcessColumns.InvoiceId.Value(row),
DateRegistered = OrderProcessColumns.DateRegistered.Value(row),
OrderNumber = OrderProcessColumns.OrderNumber.Value(row)
};
return new MappedRow<OrderProcess>(row, result);
}
}
}A MappedRow instance contains both a DataRow and the object that the DataRow mapped to.
This may be useful in situations where the DataRow contains more information than is available on the object. An example may be an OrderLine where the DataRow contains the OrderId column but the OrderLine object does not. In order to still be able to make that association it is useful to have both available.
An IAssembler implementation is used to create multiple mappings with as few calls as possible. An example may be where we perform two select queries; one to get 3 orders and another to get the order lines belonging to those 3 orders.
select OrderId, OrderNumber, OrderDate from dbo.Order where OrderId in (2, 6, 44)
| Order Id | Order Number | Order Date |
|---|---|---|
| 2 | ORD-002 | 14 Feb 2016 |
| 6 | ORD-006 | 24 Mar 2016 |
| 44 | ORD-044 | 4 Apr 2016 |
select OrderId, Product, Quantity from dbo.OrderLine where OrderId in (2, 6, 44)
| Order Id | Product | Quantity |
|---|---|---|
| 2 | Red Socks | 2 |
| 2 | Blue Socks | 3 |
| 6 | Sports Towel | 1 |
| 6 | Squash Racquet | 1 |
| 6 | Squash Ball | 3 |
| 44 | Vaughn's DDD Book | 1 |
| 44 | Shuttle.Sentinel License | 5 |
Using a MappedData instance we can keep adding the MappedRow instances to the MappedData and then have the assembler return the three Order aggregates:
public class OrderAssembler : IAssembler<Order>
{
public IEnumerable<Order> Assemble(MappedData data)
{
var result = new List<Order>();
foreach (var orderRow in data.MappedRows<Order>())
{
var order = orderRow;
foreach (var orderLineRow in data.MappedRows<OrderLine>())
{
if (orderLineRow.Row["OrderId"].Equals(order.OrderId))
{
order.AddLine(orderLineRow);
}
}
result.Add(order);
}
return result;
}
}