We Keep Coding

Mongodb collection as dynamic

I have an application that has two similar but different objects and I want to store those objects in the same collection. What is the best way to do this? And how can I query this collection?
Today my collections is represented by:
public IMongoCollection<Post> Posts
{
get
{
return _database.GetCollection<Post>("posts");
}
}
And I have this class:
public class Post
{
public string Id { get; set; }
public string Message { get; set; }
}
public class NewTypePost
{
public string Id { get; set; }
public string Image { get; set; }
}
So, today I just can save and query using Post class. Now I want to store and retrive the both classes, Post and NewTypePost.
I tried to change the class type from Post to dynamic. But when I did this, I could not query the collections.

MongoDB .NET driver offers few possibilites in such cases:
Polymorphism
You can build a hierarchy of classes and MongoDB driver will be able to determine a type of an object it gets retrieved from the database:
[BsonKnownTypes(typeof(Post), typeof(NewTypePost))]
public abstract class PostBase
{
[BsonId]
public string Id { get; set; }
}
public class Post: PostBase
{
public string Message { get; set; }
}
public class NewTypePost: PostBase
{
public string Image { get; set; }
}
MongoDB driver will create additional field _t in every document which will represent corresponding class.
Single Class
You can still have Post class and use BsonIgnoreIfNull attribute to avoid serialization exception. MongoDB .NET driver will set those properties to null if they don't exist in your database.
public class Post
{
[BsonId]
public string Id { get; set; }
[BsonIgnoreIfNull]
public string Message { get; set; }
[BsonIgnoreIfNull]
public string Image { get; set; }
}
BsonDocument
You can also drop strongly-typed approach and use BsonDocument class which is dynamic dictionary-like structure that represents your Mongo documents
var collection = db.GetCollection<BsonDocument>("posts");
More details here
dynamic
Specifying dynamic as generic parameter of ICollection you should get a list of ExpandoObject that will hold all the values you have in your database.
var collection = db.GetCollection<dynamic>("posts");
var data = collection.Find(Builders<dynamic>.Filter.Empty).ToList();
var firstMessage = data[0].Message; // dynamically typed code

Suppose I have the next conn to a test database:
var mongoClient = new MongoClient(new MongoClientSettings
{
Server = new MongoServerAddress("localhost"),
});
var database = mongoClient.GetDatabase("TestDb");
Then I can do something like:
var col = database.GetCollection<Post>("posts");
var col2 = database.GetCollection<NewTypePost>("posts");
To get two different instances of IMongoCollection but pointing to the same collection in the database. Further I am able to save to each collection in the usual way:
col.InsertOne(new Post { Message = "m1" });
col2.InsertOne(new NewTypePost { Image = "im1" });
Then, I'm also able to query from those collection base on the specific fields:
var p1= col.Find(Builders<Post>.Filter.Eq(x=>x.Message, "m1")).FirstOrDefault();
var p2 =col2.Find(Builders<NewTypePost>.Filter.Eq(x=>x.Image, "im1")).FirstOrDefault();
Console.WriteLine(p1?.Message); // m1
Console.WriteLine(p2?.Image); // im1
I don't know if that's what you want but it uses the same collection. BTW, change the Id properties to be decorated with [BsonId, BsonRepresentation(BsonType.ObjectId)]. Hope it helps.

Use the BsonDocument data type. It can do all of that. BsonDocument and dynamic back and forth is very convenient.
public class CustomObject{
public long Id{get;set;}
public string Name{get;set;}
public List<(string,object)> CollectionDynamic{get;set;}
}
// inserted in mongo
//public class CustomObject_in_Db{
// public long Id {get;set;}
// public string Name {get;set;}
// public string field2 {get;set;}
// public string field3 {get;set;}
// public string field4 {get;set;}
// public string field5 {get;set;}
// }
// something code... mapper(config)
Automapper.Mapper.CreateMap<BsonDocument,CustomObject>()
.ForMember(dest=>dest.Id, a=>a.MapFrom(s=>s.Id.GetValue(nameof(CustomObject.Id)).AsInt64)
.ForMember(dest=>dest.Name, a=>a.MapFrom(s=>s.Id.GetValue(nameof(CustomObject.Name)).AsString)
.ForMember(dest=>dest.CollectionDynamic, a=>a.MapFrom(s=>_getList(s));
// .......
private List<(string, object)> _getList(BsonDocument source){
return source.Elements.Where(e=>!typeof(CustomObject).GetProperties().Select(s=>s.Name).Any(a=>a ==e.Name)).Select(e=>e.Name, BsonTryMapper.MapToDotNetValue(e.Value)));
}

How to convert string array property to string to save in the sql server as comma seperated row mongo db to sql server C#

I have a collection in mongo db "computers",I want to retrieve the collect and save it into sql server using Entity framework. The problem is that the collection have an array object.I want to convert it to string and save. please help, Thank you. SQL server skips the string [] credentials and inserts rest.
Mongo db collection
{
"_id": ObjectId('57852fcsdsdsdsd2662a0ce400'),
"machineName": "AAADESKTOP",
"updated": ISODate('2017-05-17T15:09:39.399Z'),
"__v": 111,
"credentials": [
"####################",
"####################",
"#################",
"#####################"
],
"clientVersion": "2.12.2",
"lastActivity": ISODate('2017-05-17T16:30:50.165Z'),
"secret": "####################"
}
**C# code**
var client = new MongoClient();
var db = client.GetDatabase("Computerdata");
var collection = db.GetCollection<Computers>("computers");
var data = collection.Find(new BsonDocument()).ToListAsync().Result;
foreach (var item in data)
{
EntityFramewrokContext.Computers.Add(item);
EntityFramewrokContext.SaveChanges();
}
public class Computers
{
[BsonId]
[BsonRepresentation(BsonType.ObjectId)]
[Key]
public string Id { get; set; }
public string machineName { get; set; }
[BsonDateTimeOptions(Representation = BsonType.DateTime)]
public DateTime updated { get; set; }
public int version { get; set; }
public string[] credentials { get; set; }
public string clientVersion { get; set; }
[BsonDateTimeOptions(Representation = BsonType.DateTime)]
public DateTime lastActivity { get; set; }
public string secretKey { get; set; }
}

If you have a table called Computer in SQL Server (That you want to write these records to), then you should also create a table called ComputerCredential that has the following columns:
CredentialId int
ComputerId int
Credential nvarchar(200)
Then for each credential in the array you should create a ComputerCredential entity.
If you desperately want to do it the way you asked. Then:
var creds = String.Join(",", computer.credentials);
That will take the array of credentials and make them into a single comma separated string.
When you retrieve the values you can do this:
var creds = commaSepCredentials.Split(',');
To retrieve the array. But don't do this, do the first option and learn about relations and joins in SQL Server. It's the power in relational databases.

MongoDB best practice for referencing

I'm wondering what the best practice for modelling by using references would be given situation under. I'm using MongoRepository library.
public class User : Entity
{
publis string Id { get; set; }
public string Email { get; set; }
public string Password { get; set; }
}
public class Post : Entity
{
public string Id { get; set; }
public string Title { get; set; }
public string Summary { get; set; }
public DateTime Added { get; set; }
public User Owner { get; set; }
}
When storing the Post I want only reference to Owner (User) object instead of whole object underlying.
Currently I'm doing it like this, not knowing of better way...
var post = new Post
{
Title = "Example title",
Summary = "asd asd",
Added = DateTime.Now,
Owner = new User { Id = "someExistingUserId" }
};
postRepository.Update(post); //Save
..
//Then to get the post
var post = postRepository.GetById("previouslySavedPostId");
post.Owner = userRepository.GetById(post.Owner.Id);
return post;
userRepository and postRepository are of MongoRepository type.
Is this the correct approach to solving my problem using MongoDB with C#/MVC(4)?

You can use MongoDBRef object instead of User object.
public class Post : Entity
{
public string Id { get; set; }
public string Title { get; set; }
public string Summary { get; set; }
public DateTime Added { get; set; }
public MongoDBRef Owner { get; set; }
}
Then you can:
var mongo = new Mongo(config.BuildConfiguration());
mongo.Connect();
var DB = mongo.GetDatabase(_dataBaseName)
var post = new Post();
post.Owner = new MongoDBRef("User", userId); // First parameter is a mongoDB collection name and second is object id
// To fetch object referenced by DBRef you should do following
var owner = DB.FollowReference<User>(post.Owner);

Mongo is a document database and if you are used to using sql server it requires a slightly different way of thinking.
As you don't want the user password details in every single post, the way i would probably do it is to create a new class to contain any user info that might be required to display a post.
public class PostOwnerInfo
{
public string UserId { get; set; }
public string Name { get; set; }
}
Update your post entity, replacing the Owner property with an OwnerInfo property, of type PostOwnerInfo.
Then when you create a new post, do the following.
var user = userRepository.GetById(someExistingUserId);
var post = new Post
{
Title = "Example title",
Summary = "Example summary",
Added = DateTime.Now,
OwnerInfo = new PostOwnerInfo
{
UserId = user.Id,
Name = user.Name
}
};
postRepository.Update(post);
This way when you query for a post it will have all the user info that you require to display the post in it's OwnerInfo property with no further queries required.
var post = postRepository.GetById(previouslySavedPostId);
// post.OwnerInfo will contain user info
There is a certain amount of data redundancy, but in an document database this is how i would do it.
If you need the full user info for any reason just do a seperate query for it as you were doing before.
The idea is that you store all the user info you need for a post in a child document of the post, so you shouldn't need to do a seperate query for the user.
If the user data chages, just update the UserInfo field on all posts made by your user.
Your user data will rarely change, but you will query for posts very often.

Reuse index transformer expressions fails on Id

I'm looking to be able to reuse some of the transform expressions from indexes so I can perform identical transformations in my service layer when the document is already available.
For example, whether it's by a query or by transforming an existing document at the service layer, I want to produce a ViewModel object with this shape:
public class ClientBrief
{
public int Id { get; set; }
public string FullName { get; set; }
public string Email { get; set; }
// ellided
}
From this document model:
public class Client
{
public int Id { get; private set; }
public CompleteName Name { get; private set; }
public Dictionary<EmailAddressKey, EmailAddress> Emails { get; private set; }
// ellided
}
public class CompleteName
{
public string Title { get; set; }
public string GivenName { get; set; }
public string MiddleName { get; set; }
public string Initials { get; set; }
public string Surname { get; set; }
public string Suffix { get; set; }
public string FullName { get; set; }
}
public enum EmailAddressKey
{
EmailAddress1,
EmailAddress2,
EmailAddress3
}
public class EmailAddress
{
public string Address { get; set; }
public string Name { get; set; }
public string RoutingType { get; set; }
}
I have an expression to transform a full Client document to a ClientBrief view model:
static Expression<Func<IClientSideDatabase, Client, ClientBrief>> ClientBrief = (db, client) =>
new ClientBrief
{
Id = client.Id,
FullName = client.Name.FullName,
Email = client.Emails.Select(x => x.Value.Address).FirstOrDefault()
// ellided
};
This expression is then manipulated using an expression visitor so it can be used as the TransformResults property of an index (Client_Search) which, once it has been generated at application startup, has the following definition in Raven Studio:
Map:
docs.Clients.Select(client => new {
Query = new object[] {
client.Name.FullName,
client.Emails.SelectMany(x => x.Value.Address.Split(new char[] {
'#'
})) // ellided
}
})
(The Query field is analysed.)
Transform:
results.Select(result => new {
result = result,
client = Database.Load(result.Id.ToString())
}).Select(this0 => new {
Id = this0.client.__document_id,
FullName = this0.client.Name.FullName,
Email = DynamicEnumerable.FirstOrDefault(this0.client.Emails.Select(x => x.Value.Address))
})
However, the transformation expression used to create the index can then also be used in the service layer locally when I already have a Client document:
var brief = ClientBrief.Compile().Invoke(null, client);
It allows me to only have to have one piece of code that understands the mapping from Client to ClientBrief, whether that code is running in the database or the client app. It all seems to work ok, except the query results all have an Id of 0.
How can I get the Id property (integer) properly populated in the query?
I've read a number of similar questions here but none of the suggested answers seem to work. (Changing the Ids to strings from integers is not an option.)

I have a hard time following your sample fully, Really the best way to dig in to this would be with a failing self-contained unit test.
Nonetheless, let's see if I can pull out the important bits.
In the transform, you have two areas where you are working with the id:
...
client = Database.Load(result.Id.ToString())
...
Id = this0.client.__document_id,
...
The result.Id in the first line and the Id = in the second line are expected to be integers.
The Database.Load() expects a string document key and that is also what you see in __document_id.
The confusion comes from Raven's documentation, code, and examples all use the terms id and key interchangeably, but this is only true when you use string identifiers. When you use non-string identifiers, such as ints or guids, the id may be 123, but the document key is still clients/123.
So try changing your transform so it translates:
...
client = Database.Load("clients/" + result.Id)
...
Id = int.Parse(this0.client.__document_id.Split("/")[1]),
...
... or whatever the c# equivalent linq form would be.

Entity Framework (5.0) Code First - Insert into Collection within Collection

I've got three classes.
Event > Workshop > Workshop Times
I'm currently looking for best way of inserting records into the Workshop Times, this is running through code first using ICollections.
Looking for something along the lines of this, but I know it doesn't work:
//Create connection
var db = new Context();
var Event = db.Events
.Include("Workshops")
.Include("Workshops.Times")
.Where(ev => ev.GUID == EventGUID).FirstOrDefault();
Event.Workshops.Add(new Workshop
{
Name = tbWorkshopName.Text,
Description = tbWorkshopDescription.Text,
Times.Add(new WorkshopTime{
//Information for times
})
});
db.SaveChanges();
Chopped down classes:
public class Workshops{
public int id { get; set; }
public string name { get; set; }
public ICollection<WorkshopTimes> Times{get;set;}
}
public class Events {
public int id { get; set; }
public string name { get; set; }
public ICollection<Workshops> WorkShops { get; set; }
}
public class WorkshopTimes {
public int id { get; set; }
public DateTime time { get; set; }
}

You are definitely on the right track with your query, however your include statements appear incorrect. From your model I would expect:
var Event = db.Events
.Include("WorkShops")
.Include("WorkShops.events")
.Where(ev => ev.GUID == EventGUID).FirstOrDefault();
Note this uses the property names not the types. This will ensure that the entities in the listed nav properties will be included in the result.
In addition you can use a lambda to do the same thing (but its typesafe)
Check out here for how to do a very similar scenario to yours:
EF Code First - Include(x => x.Properties.Entity) a 1 : Many association
or from rowan miller (from EF team)
http://romiller.com/2010/07/14/ef-ctp4-tips-tricks-include-with-lambda/
And make sure you are using System.Data.Entities for lambda based includes ( Where did the overload of DbQuery.Include() go that takes a lambda? )