Skip to content
This repository has been archived by the owner on Oct 20, 2022. It is now read-only.

This projects implements an abstraction for Azure Storage Tables to use POCOs

Notifications You must be signed in to change notification settings

ecomerc/AzureStorageTable

 
 

Repository files navigation

Build status

AzureStorageTable

This projects implements an abstraction for Azure Storage Tables to use POCOs because deriving every entity from ITableEntity or TableEntity looks like a step backwards. The current implementation is intended to be an abstraction to store every existing entity into Azure Table Store.

There are two different principals implemented. The first allows to define an external mapping structure between the existing model and the required fields in Azure Table, e.g. Partition and RowKey. The second option is to decorate existing models with attributes to map the properties to partition and rowkey.

Installation

Install-Package CoreHelpers.WindowsAzure.Storage.Table

Manual Entity Mapper

// create a new user model
var user = new UserModel() { FirstName = "Egon", LastName = "Mueller", Contact = "[email protected]" };

using (var storageContext = new StorageContext(storageKey, storageSecret))
{
  // configure the entity mapper
  storageContext.AddEntityMapper(typeof(UserModel), new DynamicTableEntityMapper() { TableName = "UserProfiles", PartitionKeyPropery = "Contact", RowKeyProperty = "Contact" });

  // ensure the table exists
  storageContext.CreateTable<UserModel>();

  // inser the model
  storageContext.MergeOrInsert<UserModel>(user);

  // query all
  var result = storageContext.Query<UserModel>();

  foreach (var r in result)
  {
    Console.WriteLine(r.FirstName);
  }
}

Attribute Based Entity Mapper

Decorate your existing model

[Storable()]
public class UserModel2
{                       
  [PartitionKey]
  [RowKey]
  public string Contact { get; set; }

  public string FirstName { get; set; } 
  public string LastName { get; set; }                		
}

Configure and use the Storage Context

// create a new user model
var user = new UserModel2() { FirstName = "Egon", LastName = "Mueller", Contact = "[email protected]" };            

using (var storageContext = new StorageContext(storageKey, storageSecret))
{
  // ensure we are using the attributes
  storageContext.AddAttributeMapper();

  // ensure the table exists
  storageContext.CreateTable<UserModel2>();

  // inser the model
  storageContext.MergeOrInsert<UserModel2>(user);

  // query all
  var result = storageContext.Query<UserModel2>();

  foreach (var r in result)
  {
      Console.WriteLine(r.FirstName);
  }
}

Virtual Partition and Row-Keys

When implementing storage schemes in Azure Table sometimes the partition or the row key are combinations out for two or more properties. Because of that the Azure Storage Table components supports virtual partition and row key attributes as follows:

[Storable()]
[VirtualPartitionKey("{{Value1}}-{{Value2}}")]
[VirtualRowKey("{{Value2}}-{{Value3}}")]
public class VirtualPartKeyDemoModel
{
  public string Value1 { get; set;  }
  public string Value2 { get; set;  }				
  public string Value3 { get; set;  }
}

Virtual Array Attributes

When storing arrays in Azure Table store there are two options. The first option is to store it as a JSON payload and the second option is to expand the array with his items to separate properties, e.g.

{ DataElements: [1,2,3,4] }

becomes

DE00 DE01 DE02 DE03
1 2 3 4

in Azure Table Store with the following code:

[Storable(Tablename: "VArrayModels")]
public class VArrayModel
{
  [PartitionKey]
  [RowKey]
  public string UUID { get; set; }

  [VirtualList(PropertyFormat: "DE{{index}}", Digits: 2)]
  public List<int> DataElements { get; set; } = new List<int>();
}

Store as JSON Object Attribute

The store as JSON attribute allows to store refenrenced objects as json payload for a specific property

[Storable(Tablename: "JObjectModel")]
public class JObjectModel
{
 [PartitionKey]
 [RowKey]
 public string UUID { get; set; }
 
 [StoreAsJsonObject]
 public Dictionary<string, string> Data { get; set; } = new Dictionary<string, string>();
}

Related tables

It is possible to automatically load related tables, either lazily or eagerly. In order to load lazily simply pack the object in the Lazy<> type.

[Storable(Tablename: "JObjectModel")]
public class Model
{
 [PartitionKey]
 [RowKey]
 public string UUID { get; set; }

 public string UserId {get; set; }
 
 //This is the rowkey of the OtherModel
 public string OtherModel { get; set; } 

 //Partition key must be specified explicitly, rowkey defaults to the name of the type (here: OtherModel)
 [RelatedTable("UserId")]
 public Lazy<OtherModel> OtherModelObject { get; set; } 
}

It is possible to specify the rowkey explicitly:

[Storable(Tablename: "JObjectModel")]
public class Model
{
 [PartitionKey]
 [RowKey]
 public string UUID { get; set; }

 public string UserId {get; set; }
 
 public string OtherModelId { get; set; } 

 [RelatedTable("UserId", RowKey="OtherModelId")]
 public OtherModel OtherModel { get; set; } 
}

If neither the rowkey or the partition key is the name of a property of the object they are used directly as strings, and obviously to reduce the possible causes of errors it is recommended to use the nameof:

[Storable(Tablename: "Models")]
public class Model
{
 [PartitionKey]
 [RowKey]
 public string UUID { get; set; }
 
 public string OtherModelId { get; set; } 

 [RelatedTable(nameof(UUID), RowKey=nameof(OtherModelId))]
 public Lazy<OtherModel> OtherModel { get; set; } 
}

About

This projects implements an abstraction for Azure Storage Tables to use POCOs

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • C# 97.9%
  • PowerShell 2.1%