I have the following problem: when i try to change the town of a user by passing the modified user entry to my ModifyUser method in my UserService class everything seems ok but the changes are not applied in the database.
I tried the solutions from similar problems i found on stackoverflow but none of them seem to work.
As they say "a picture is worth a thousand words" take a look at this, since its easier to understand compared to my short explanation.
http://prntscr.com/el51z0
else if (property == "BornTown")
{
if (this.townService.TownExists(value))
{
User user = this.userService.GetUserByName(username);
Town town = this.townService.GetTownByName(value);
user.BornTown = town;
this.userService.ModifyUser(user);
}
else
{
Console.WriteLine($"Value {value} is not valid.");
throw new ArgumentException($"Town {value} not found!");
}
}
When i pass the user to the ModifyUser the database does not update, and BornTown stays NULL.
public void ModifyUser(User user)
{
using (PhotoShareContext context = new PhotoShareContext())
{
context.Users.Attach(user);
context.Entry(user).State = EntityState.Modified;
context.SaveChanges();
}
}
Or else try below code:
User user = context.users.where(x=>x.username == username).FirstOrDefault();
user.BornTown = town; //string
context.savechanges();
Can you try
context.Users.where(x=>x.username == username).FirstOrDefault();
and don't use any()
Can you check if you are getting the proper username, by again checking the username once you are getting back the value in the function.
Instead of manipulating the navigation properties directly try to manage the entity relation using it's foreign keys:
Town town = this.townService.GetTownByName(value);
user.BornTownId = town.Id;
Keep in mind that if you have not created explicit foreign keys in your model entity you should map the existing one: take a look at "Users" table - see how the foreign key column is named- it should look something like UserTown_Id. Then put it in your model class (you may add foreign key attribute or map it in model builder).
Be careful when adding the new property to your model (it can make your optional relation with "Towns" required).
Related
I am developing code for app engine. I tried to update an existing row by updating the same entity returned as result of query. But it creates new row instead of updating the same row. Following is the code:
public boolean updateProfile(DbProfile profile) {
Transaction txn = _datastore.beginTransaction();
Entity entity = getProfileEntity(profile.getLoginId());
if (entity != null) {
entity.setProperty(DbProfile.DbProfilePropertyNames.address, profile.getAddress());
entity.setProperty(DbProfile.DbProfilePropertyNames.name, profile.getName());
Key key = _datastore.put(entity);
txn.commit();
return true;
}
return false;
}
private Entity getProfileEntity(String userName) {
Key eRecommendationKey = KeyFactory.createKey("eRecommendation", _dbKeyName);
FilterPredicate predicateUsername =
new FilterPredicate(DbProfile.DbProfilePropertyNames.loginId, FilterOperator.EQUAL,
userName.toUpperCase());
Query query =
new Query(DbProfile.entityProfileName, eRecommendationKey).setFilter(predicateUsername);
List<Entity> profiles =
_datastore.prepare(query).asList(FetchOptions.Builder.withDefaults());
Utils.log.log( Level.SEVERE, "not found"+profiles.size() );
if (profiles.size() == 0) {
//profile data is not set yet
return null;
} else {
return profiles.get(0);
}
}
Following image shows fields in the entity.
Please let me know how can I fix the issue.
My Java skills are not too good, so I find it difficult to understand your code sample. I also don't see where updateProfile() is called and how your code is getting or constructing the profile object, especially whether the key of the profile object is altered.
But in general, if new entities are created instead of updating existing entities, the reason is that the key at your updating commit is different from the actual key of the existing entity.
Typical situations:
a string (key-name) is used instead of an integer (ID), or vice versa
a typo in the kind name of the key
different namespace or app properties of the key
parents are missing in the key path or are constructed wrongly
Suggestion:
In datastore viewer, compare the key of an existing entity with the key of the accidentally created entity. The difference between both keys might give you a hint where to look in your code for the bug.
I solved the problem. It was my mistake. I had called saveData servlet instead of updateProfile servlet.
TL;DR What is the proper way of rehydrating an entity framework object with a self referential many to many relationship from a DTO and updating it with the new values so that the database updates correctly?
I have the following entity (irrelevant stuff trimmed)
public class Role
{
[Key]
[Required]
public String RoleId { get; set; }
public List<Role> Children { get; set; }
}
In my dbContext, I have set up a many to many relationship
modelBuilder.Entity<Role>().HasMany(r => r.Children).WithMany();
I'm using MVC front end, with a web-api backend for an n-tier setup, and an mssql database.
The following chain of events happens
Browser->MVC Controller->REST call to Web API->WebAPI Controller->DB Context Query
This chain happens twice, once to view the page in edit mode, and then again when the user pushes the save button to persist.
When setting children on the entity, they always already exist first (IE, you don't create the parent and the children at the same time, you are just adding an existing child to a parent)
There is a DTO used by the MVC model and web API, which I re-hydrate to the entity on the web-api side.
public IHttpActionResult UpdateRoleInfo(RoleVM roleInfo){
//lookup existing entity to update
var existing = db.Roles.FirstOrDefault(y => y.RoleId == roleInfo.ExistingRoleId);
...Something happens here (see below for things i've tried)...
db.SaveChanges();
}
My first try was this :
existing.Children = roleInfo.Children
This tried to recreate all of the existing children as part of the save. (Primary key constraint violation on the roles table)
I changed that to
//Fetch all of the roles from the database to lookup the existing children
var allRoles = GetRoles();
//Have to reselect the roles from the DB so the DB doesn't try to recreate new ones for the children.
var childrenToAdd = roleInfo.Roles.Select(role2 => allRoles.FirstOrDefault(r => r.RoleId == role2.RoleId)).ToList();
existing.Children = childrenToAdd;
This correctly works for updating a role that does not already have any children, to add some the first time, but if you update a role that already has children, it tries to re-add the children to the database a second time, getting a primary key violation on the roles_role table
I then tried pre-pending this code to the second one above,
existing.Children.Clear();
db.SaveChanges();
I would expect this to delete all the existing parent-child relationships from the many to many table for this parent, and then recreate them with the new children. Why not?
TL;DR What is the proper way of rehydrating an entity framework object with a self referential many to many relationship from a DTO and updating it with the new values so that the database updates correctly?
Try turning off auto detect changes (before retrieving from the DB) via
context.Configuration.AutoDetectChangesEnabled = false;
Then set the state to modified on the specific role object you are updating
context.Entry(role).State = EntityState.Modified;
Haven't tried this myself on a self-referencing many-to-many table, but adding & updating entities in the manner can save all sorts of headaches where EF incorrectly infers what you are adding/updating
Found the problem.
On the initial load of the entity, I was using an include statement to eager load the children.
When I updated the entity, when I fetched it from the db again, I did not eager load the children. Therefore the additions/updates were getting confused. Once I put the include in during the upload Scenario #2 above worked (the explicit clear was not needed)
db.Roles.Include("Children").FirstOrDefault(z => z.RoleId == RoleId);
Also related, if you have this same problem when dealing with relationships across different tables, make sure all the entities that are involved in the graph are from the same DB context!
https://msdn.microsoft.com/en-us/magazine/dn166926.aspx
I'm currently developing a mobile application who uses a Google App Engine-hosted web service.
But i'm facing an issue. I just want to add a field in one my database's table.
App Engine doesn't use classic SQL syntax, but GQL. So i cannot use the ALTER TABLE statement. How can i do this with GQL ? I looked for a solution on the web, but there's not a lot of help.
public MyEntity() {
}
#Id
#GeneratedValue(strategy=GenerationType.IDENTITY)
private Key idStation;
private String name;
private double longitude;
private double latitude;
private java.util.Date dateRefresh = new Date(); //the field i want to add in DB
So, now when i create a "MyEntity" object, it should add the "dateRefresh" field into the database... I create my object like this:
MyEntity station = new MyEntity();
station.setName("test");
station.setLatitude(0);
station.setLongitude(0);
station.setDateRefresh(new Date("01/01/1980"));
DaoFactory.getStationDao().addStation(station);
addStation method:
#Override
public MyEntity addStation(MyEntity station) {
EntityManager em = PersistenceManager.getEntityManagerFactory().createEntityManager();
try {
em.getTransaction().begin();
em.persist(station);
em.getTransaction().commit();
} finally {
if(em.getTransaction().isActive()) em.getTransaction().rollback();
em.close();
}
return station;
}
The field "dateRefresh" is never created into my DB...
Someone to help me please ?
Thanks in advance
Just add another field to your data structure, maybe providing a default clause, and that's all. For example, if you have a UserAccount:
class UserAccount(db.Model):
user = db.UserProperty()
user_id = db.StringProperty()
you may easily add:
class UserAccount(db.Model):
user = db.UserProperty()
user_id = db.StringProperty()
extra_info = db.IntegerProperty(default=0)
timezone = db.StringProperty(default="UTC")
and let it go.
While the datastore kinda mimics tables, data is stored on a per entity basis. There is no schema or table.
All you need to do is update your model class, and new entities will be saved with the structure (fields) of the new entity.
Old entities and indexes, however, are not automatically updated. They still have the same fields as they had when they were originally written to the datastore.
There's two ways to do this. One is to make sure your code can handle situations where your new properties are missing, ie make sure no exceptions are thrown, or handle the exceptions properly when you're missing the properties.
The second way is to write a little function (usu a mapreduce function) to update every entity with appropriate or null values for your new properties.
Note that indexes are not updated unless the entity is written. So if you add a new indexed property, old entities won't show up when you query for the new property. In this case, you must use the second method and update all the entities in the datastore so that they are indexed.
I've got an application that reads Lead records from Salesforce via the API and I want to link the Lead Owner field to an attribute in the application. The Lead Owner field doesn't up in the list of available fields but all the custom fields do.
So, my first attempt at a solution was to create a custom field that displayed the Lead Owner name. In the SF formula editor, as far as I can tell, it doesn't display the actual data field but instead displays the ID string. Which is pretty meaningless in the context that I need it for.
alt text http://skinny.fire-storm.net/forposting/insertfield.JPG
Is there a way that we can get at the data in the object that the ID string references?
alt text http://skinny.fire-storm.net/forposting/havewant.JPG
I have the RED BOX but need the GREEN BOX.
EDIT: I can't change the application code that calls the API. I can only change salesforce. So, this is more of a salesforce superuser / formula-writer question, not a question about writing code that calls the SF API.
Salesforce allows access to related data through what they call relationship queries. Instead of joining, you specify the query like this:
System.debug([SELECT Owner.Name FROM Lead WHERE Id = '00QS00000037lvv'].Owner.Name);
Try running that in the system log, just replace the lead ID with one that you're looking at.
When accessing the data through the API, the principle is the same, your proxy objects should allow you to access Lead.Owner.Name.
EDIT:
I agree with eyescream, since you can't change the application code, creating an Apex trigger would be the best way to go here. Here's some code to get you started:
trigger Lead_UpdateOwner on Lead(before insert, before update)
{
Map<Id, String> ownerMap = new Map<Id, String>();
for (Lead lead : Trigger.new)
{
ownerMap.put(lead.OwnerId, null);
}
if (ownerMap.size() > 0)
{
for (User[] users : [SELECT Id, Name FROM User WHERE Id IN :ownerMap.keySet()])
{
for (Integer i=0; i<users.size(); i++)
{
ownerMap.put(users[i].Id, users[i].Name);
}
}
for (Lead lead : Trigger.new)
{
lead.OwnerName__c = ownerMap.get(lead.OwnerId);
}
}
}
lead.OwnerName__c would need to be the name of your custom field on the lead object that will hold the owner name. Type Text, length 121.
I had a similar problem, but wanted all the current and future User fields available. Since a custom lookup field to the User is not restricted by formula fields, I created one named
OwnerLookup
on the Opportunity and Account objects, then used a triggers to populate it on creation or edit. For example the Opportunity trigger is this:
trigger OpportunityTrigger on Opportunity (before insert, after insert, before update, after update) {
if(trigger.isBefore && trigger.isInsert) {
OpportunityTriggerHandler.newOpportunity(Trigger.old, Trigger.new);
}
else if(trigger.isAfter && trigger.isInsert){
//OpportunityTriggerHandler.futureUse(Trigger.new);
}
else if(trigger.isBefore && trigger.isUpdate){
OpportunityTriggerHandler.updateOpportunity(Trigger.new, Trigger.oldMap);
}
else if(trigger.isAfter && trigger.isUpdate){
//OpportunityTriggerHandler.futureUse(Trigger.new, Trigger.oldMap);
}
}
and the OpportunityTriggerHandler class (Apex code) is:
public with sharing class OpportunityTriggerHandler {
public static void newOpportunity( List<Opportunity> oldOpportunitys, List<Opportunity> newOpportunitys ) {
for (Opportunity opp: newOpportunitys) {
updateOwnerData( opp );
}
}
public static void updateOpportunity( List<Opportunity> oldOpportunitys, Map<Id, Opportunity> newOpportunitys ) {
for (Opportunity opp: oldOpportunitys) {
updateOwnerData( opp );
}
}
public static void updateOwnerData( Opportunity opp ) {
opp.OwnerLookup__c = opp.OwnerId;
}
}
I then create Formula fields on the Opportunity/Account objects to get to any of the owner (User) object fields, such as Oppty Owner Name formula field:
OwnerLookup__r.FirstName & " " & OwnerLookup__r.LastName
VLOOKUP function would be a good try, but
it's available only in validation rules, not in field definitions
it can be used only on custom objects and you need data from User
I'd say you need to query from your application for
SELECT Owner.FirstName, Owner.LastName FROM Lead
Other than that... some "after insert, after update" trigger that would populate your custom field when owner changes?
Just posting for completeness sake (and for the Google searches): The issue arises with any object that can be queued, not just Lead, since the source of it is that the owner can refer to either a user (as usual) or a queue object.
This can be resolved using a formula field instead of triggers, like below:
BLANKVALUE(Owner:Queue.QueueName, Owner:User.FirstName & " " & Owner:User.LastName)
Basically, the BLANKVALUE function in the formula checks whether the owner.queuename is blank, and if so gives the name of the user.
Today I've got a problem when I tried using following code to alter the model attribute in the controller
function userlist($trigger = 1)
{
if($trigger == 1)
{
$this->User->useTable = 'betausers'; //'betausers' is completely the same structure as table 'users'
}
$users = $this->User->find('all');
debug($users);
}
And the model file is
class User extends AppModel
{
var $name = "User";
//var $useTable = 'betausers';
function beforeFind() //only for debug
{
debug($this->useTable);
}
}
The debug message in the model showed the userTable attribute had been changed to betausers.And It was supposed to show all records in table betausers.However,I still got the data in the users,which quite confused me.And I hope someone can show me some directions to solve this problem.
Regards
Model::useTable is only consulted during model instantiation (see the API documentation for Model::__construct). If you want to change the model's table on the fly, you should use Model::setSource:
if ( $trigger == 1 ) {
$this->User->setSource('betausers');
}
The table to use is "fixed" when the model is loaded/instantiated. At that time a DB connection object is created, the table schema is being checked and a lot of other things happen. You can change that variable later all you want, Cake is not looking at it anymore after that point.
One model should be associated with one table, and that association shouldn't change during runtime. You'll need to make another model BetaUser and dynamically change the model you're using. Or rethink your database schema, a simple flag to distinguish beta users from regular users within the users table may be better than a whole new table.