I have a scenario where I have to update an entity if it exists or add a new one if it doesn't.
I would like to execute a single method for this (it would be great if it were a single trip to the server).
Is there something like that in EF?
Right now my code looks like this:
var entity = db.Entities.FirstOrDefault(e => e.Id == myId);
if (entity == null)
{
entity = db.Entities.CreateObject();
entity.Id = myId;
}
entity.Value = "my modified value";
db.SaveChanges();
But I would like to avoid the first query, something like this:
var entity = new Entity();
entity.Id = myId;
entity.Value = "my modified value";
db.AddOrAttach(entity);
db.SaveChanges();
Is there anything similar? or do I have to perform the first query no matter what?
Thanks
You have to perform the first query no matter what, unfortunately.
One option would be to write a stored procedure that performs a T-SQL MERGE then map it to a function import, though that would require that you pass the scalar values of the entity as parameters (and support for navigation properties would be done), but it would accomplish what you're after.
I ran some quick test code for editing in MVC 3 with EF 4 and it seems to work for edit with following code:
using (var context = new TestStackOverFlowEntities())
{
Person p = new Person();
p.Id = long.Parse(collection["Id"]);
p.FirstName = collection["FirstName"];
p.LastName = collection["LastName"];
context.People.Attach(p);
context.ObjectStateManager.ChangeObjectState(p, System.Data.EntityState.Modified);
context.SaveChanges();
return RedirectToAction("Index");
}
Edit: I checked with creating new object too, you need to change this
context.ObjectStateManager.ChangeObjectState(p, System.Data.EntityState.Added);
when Id == 0 //ie new object.
Quick and dirty code to add new is this:
using (var context = new TestStackOverFlowEntities())
{
Person p = new Person();
p.Id = 0;
p.FirstName = collection["FirstName"];
p.LastName = collection["LastName"];
context.People.Attach(p);
context.ObjectStateManager.ChangeObjectState(p, System.Data.EntityState.Added);
context.SaveChanges();
return RedirectToAction("Index");
}
If you are just trying to limit code to clarify your controllers:
db.Attach(model);
db.SaveChanges(model);
Will update if the Entity Key exists, and create if it does not.
Related
I am new to elasticsearch so I will need some help. Unfortunately, I didnt found the answer in other topics here on SO.
I have some .net core application which I inherited and now there is a need to implement some changes.
I already have a method of getting data from elasticsearch, but after getting them, I am not sure how to change it and use it in application.
To be precise, I need to parse first and last name and to remove special characters, specific serbian latin letters like "šđžčć" etc... I already have a method for this parsing written but not sure how to call it...
So, my question is can I and how can I do this?
What I have now is the following:
var result = await _elasticClient.SearchAsync<CachedUserEntity>(
s =>
s.Index(_aliasName)
.Query(q => andQuery));
CachedUserEntity, among others, contains property about FirstName and LastName.
Inside results.Documents, I am getting the data about FirstName and LastName from elasticsearch, but I am not sure how to access it in order to update it via aformentioned NameParser ...
Sorry if the question is too easy, not to say stupid :)
I wont use updateByQuery here, for some reasons. I would scroll on documents (i use matchAll on my exemple, you obviously need to replace it with your query), or, if you dont know how to identify documents to update, only update usefull documents in UpdateManyWithIndex/UpdateManyPartial function.
For performance, we have to update severals documents at once, so we use bulk/updateMany function.
You can use both solution, the classic update, or the second (partial update) with an object containing the targeteds fields.
On server sides, both solutions will have the same cost / performance.
var searchResponse = Client.Search<CachedUserEntity>(s => s
.Query(q => q
MatchAll()
)
.Scroll("10s")
);
while (searchResponse.Documents.Any())
{
List<CachedUserEntity> NewSearchResponse = RemoveChar(searchResponse);
UpdateManyWithIndex<CachedUserEntity>(NewSearchResponse, _aliasName);
searchResponse = Client.Scroll<Project>("2h", searchResponse.ScrollId);
}
public void UpdateManyWithIndex<C>(List<C> obj, string index) where C : class {
var bulkResponse = Client.Bulk(b => b
.Index(index).Refresh(Elasticsearch.Net.Refresh.WaitFor) // explicitly provide index name
.UpdateMany<C>(obj, (bu, d) => bu.Doc(d)));
}
Or, using partial update object
Note: in this case Indix is already set on my client (add .index if needed)
var searchResponse = Client.Search<CachedUserEntity>(s => s
.Query(q => q
MatchAll()
)
.Scroll("2h")
);
while (searchResponse.Documents.Any())
{
List<object> listPocoPartialObj = GetPocoPartialObjList(searchResponse);
UpdateManyPartial(listPocoPartialObj);
searchResponse = Client.Scroll<Project>("2h", searchResponse.ScrollId);
}
private List<object> GetPocoPartialObjList(List<CachedUserEntity> cachedList) {
List<object> listPoco = new List<object>();
//note if you dont have cachedList.Id, take a look at result.source, comments if needed
foreach (var eltCached in cachedList) {
listPoco.Add( new object() { Id = cachedList.Id, FirstName = YOURFIELDWITHOUTSPECIALCHAR, LastName = YOURSECONDFIELDWITHOUTSPECIALCHAR});
}
return listPoco;
}
public bool UpdateManyPartial(List<object> partialObj)
{
var bulkResponse = Client.Bulk(b => b
.Refresh(Elasticsearch.Net.Refresh.WaitFor)
.UpdateMany(partialObj, (bu, d) => bu.Doc(d))
);
if (!bulkResponse.IsValid)
{
GetErrorMsgs(bulkResponse);
}
return (bulkResponse?.IsValid == true);
}
_context.Update(v) ;
_context.SaveChanges();
When I use this code then SQL Server adds a new record instead of updating the
current context
[HttpPost]
public IActionResult PageVote(List<string> Sar)
{
string name_voter = ViewBag.getValue = TempData["Namevalue"];
int count = 0;
foreach (var item in Sar)
{
count = count + 1;
}
if (count == 6)
{
Vote v = new Vote()
{
VoteSarparast1 = Sar[0],
VoteSarparast2 = Sar[1],
VoteSarparast3 = Sar[2],
VoteSarparast4 = Sar[3],
VoteSarparast5 = Sar[4],
VoteSarparast6 = Sar[5],
};
var voter = _context.Votes.FirstOrDefault(u => u.Voter == name_voter && u.IsVoted == true);
if (voter == null)
{
v.IsVoted = true;
v.Voter = name_voter;
_context.Add(v);
_context.SaveChanges();
ViewBag.Greeting = "رای شما با موفقیت ثبت شد";
return RedirectToAction(nameof(end));
}
v.IsVoted = true;
v.Voter = name_voter;
_context.Update(v);
_context.SaveChanges();
return RedirectToAction(nameof(end));
}
else
{
return View(_context.Applicants.ToList());
}
}
You need to tell the DbContext about your entity. If you do var vote = new Vote() vote has no Id. The DbContext see this and thinks you want to Add a new entity, so it simply does that. The DbContext tracks all the entities that you load from it, but since this is just a new instance, it has no idea about it.
To actually perform an update, you have two options:
1 - Load the Vote from the database in some way; If you get an Id, use that to find it.
// Loads the current vote by its id (or whatever other field..)
var existingVote = context.Votes.Single(p => p.Id == id_from_param);
// Perform the changes you want..
existingVote.SomeField = "NewValue";
// Then call save normally.
context.SaveChanges();
2 - Or if you don't want to load it from Db, you have to manually tell the DbContext what to do:
// create a new "vote"...
var vote = new Vote
{
// Since it's an update, you must have the Id somehow.. so you must set it manually
Id = id_from_param,
// do the changes you want. Be careful, because this can cause data loss!
SomeField = "NewValue"
};
// This is you telling the DbContext: Hey, I control this entity.
// I know it exists in the DB and it's modified
context.Entry(vote).State = EntityState.Modified;
// Then call save normally.
context.SaveChanges();
Either of those two approaches should fix your issue, but I suggest you read a little bit more about how Entity Framework works. This is crucial for the success (and performance) of your apps. Especially option 2 above can cause many many issues. There's a reason why the DbContext keep track of entities, so you don't have to. It's very complicated and things can go south fast.
Some links for you:
ChangeTracker in Entity Framework Core
Working with Disconnected Entity Graph in Entity Framework Core
I am working on Database First Approach in Entity Framework where I have to retrieve specific columns from the Entity.
Public IQueryable<Entity.Employees> GetEmployeeName(String FName,String LName)
{
var query = (from s in Employees
where s.firstName = FName && s.lastName = LName
select new {s.firstName, s.middleName});
return query;
}
Here return statement is throwing an error where it seems that its not matching with Employees (entity) columns. Could you please help me in sorting out this issue? Thanks in advance.
You need to use == for comparison, also you need to use dynamic type as return type since you are returning a custom anonymous type. Try this
Public IQueryable<dynamic> GetEmployeeName(String FName,String LName)
{
var query=(from s in Employees
where s.firstName==FName && s.lastName==LName
select new {s.firstName,s.middleName});
return query.AsQueryable();
}
Finally you will use it like below, keep in mind that intelisense won't work on dynamic object.
var query = GetEmployeeName("Jake", "Smith");
List<dynamic> results = query.ToList();
foreach (dynamic result in results)
{
string fristName = result.FirstName;
string lastName = result.MiddleName;
}
I have the following method which is called from Ajax:
[Authorize]
[ValidateInput(false)]
[ScriptMethod(ResponseFormat = ResponseFormat.Json)]
public JsonNetResult CreateOrUpdateTimeRecord(TimeRecord tr)
{
TimeRecord trLocal;
if (tr.Id == -1 || tr.Id == 0)
{
trLocal = new TimeRecord
{
Description = tr.Description,
StartTime = tr.StartTime,
EndTime = tr.EndTime,
User =new myTimeMvc.Models.NHibernate.Models.User {Id = tr.User.Id},// _userRepo.Get(tr.User.Id),
Hdt = new Hdt {Id = tr.Hdt.Id}//_hdtRepo.Get(tr.Hdt.Id)
};
_timeRepo.Insert(trLocal);
}
else
{
trLocal = _timeRepo.Get(tr.Id);
trLocal.Description = tr.Description;
trLocal.StartTime = tr.StartTime;
trLocal.EndTime = tr.EndTime;
_timeRepo.Update(trLocal);
}
...
}
As you can see my TimeRecord has a reference to User and Hdt. Now I started to work with NHibernate Profiler which complains when I resolve my properties by loading them from their coresponding repositories. Which is clear to me since I actually don't need to query the database for that since I have the ID's for this objects.
User = _userRepo.Get(tr.User.Id),
Hdt = _hdtRepo.Get(tr.Hdt.Id)
But I'm not 100% sure if I can use this instead:
User =new myTimeMvc.Models.NHibernate.Models.User {Id = tr.User.Id},,
Hdt = new Hdt {Id = tr.Hdt.Id}
I guess NHibernate lazy proxies work the same way since they only contain just the ID of the related object and load the rest when it is needed. Do I have to attach this "new" oject anyway to my session?
Can someone tell me what is the correct way to do this?
Cheers,
Stefan
There are a few ways how to achieve that. One of them could be using the Load() method. Check Ayendes post: NHibernate – The difference between Get, Load and querying by id, an extract:
Load will never return null. It will always return an entity or throw an exception. Because that is the contract that we have we it, it is permissible for Load to not hit the database when you call it, it is free to return a proxy instead.
Other words, we can do something like this
User = _userRepo.Load(tr.User.Id),
Hdt = _hdtRepo.Load(tr.Hdt.Id)
Where the Load would be encapsulating the session.Load()
Am developing a ViewModel/PresentationModel which is getting complex.
I want the Linq query to return an IQueryable<UserPresentationModel>
Using EntityFramework against MSSQL
Is it possible to do any sort of iteration over the set before returning it to the presentation layer ie
List<UserPresentationModel> list = new List<UserPresentationModel>();
foreach (var person in listOfPeople)
{
UserPresentationModel u = new UserPresentationModel();
int userUIStatus = GetColourStateOfPerson(person);
u.FirstName = person.FirstName;
u.UserUIStatus = userUIStatus;
list.Add(u);
}
return list
This feels like it would always be N+1, and I'd never get the advantages of deferred execution, composing of queries..
Or (and I think am answering my own question) do I need to think in a SQL set based manner.
First, we can convert your code to LINQ.
IEnumerable<UserPresentationModel> models =
from person in listOfPeople
select new UserPresentationModel
{
FirstName = person.FirstName,
UserUIStatus = GetColourStateOfPerson(person)
}
return models.ToList();
Now, if GetColourStateOfPerson is making a DB round-trip, you definitely want to pull that out.
IDictionary<int, int> colourStatesByPersonId = GetColourStatesOfPeople(listOfPeople);
IEnumerable<UserPresentationModel> models =
from person in listOfPeople
select new UserPresentationModel
{
FirstName = person.FirstName,
UserUIStatus = colourStatesByPersonId[person.PersonId]
}
return models.ToList();
You could probably manage to create a single LINQ query that grabs just the first names and colour states of the people you want in a single query, but you haven't provided enough information about your data context for me to help you with that.
I would personally avoid passing around an IQueryable, which could continue making database trips any time somebody touches it. Let your data layer get out all the data you're likely to need, compose it into a list, and return that.
use IEnumerable<T>.Aggregate() instead of looping.
return listOfPeople.Aggregate(new List<UserPresentationModel>(), person => {
return new UserPresentationModel {
FirstName = person.FirstName,
UserUIStatus = GetColourStateOfPerson(person)
};
}).AsQueryable();
return listOfPeople.AsEnumerable().Select(p =>
new UserPresentationModel
{
FirstName = p.FirstName,
UserUIStatus = GetColourStateOfPerson(p)
}).AsQueryable();
I'm assuming that listOfPeople is an IQueryable that will eventually execute against your database. If that is the case then AsEnumerable() is important because SQL Server won't know what to do with GetColourStateOfPerson(). AsEnumerable() will force the IQueryable's expression tree to execute, pull the resulting rows out of your database and then apply Select() transformation in code as oppose to in SQL Server.
If you can implement GetColourStateOfPerson() as a stored proc or database function then you can omit AsEnumerable() and AsQueryable() and allow execution to delay even longer.