I have created a set of T4 templates to generate the unit classes for a DAL. An issue I have is with providing a string of the correct length based on "StringLength" attribute on the properties of the entities.
I tried accessing CodeProperty.Attributes, but it says there are zero attributes where I know that to be incorrect.
Any help would be greatly appreciated!
Just in case someone else stumbles upon this......
I was unable to utilize CodeProperty to obtain the attribute. Instead I use the following given I already have a CodeClass and CodeProperty to use in my search:
private int GetStringLength(CodeProperty codeProperty, CodeClass codeClass)
{
var type = Type.GetType(codeClass.Name); //you may need to use assembly qualified name
var props = type.GetProperties();
if (props != null && props.Any(p => p.Name.Equals(codeProperty.Name)))
{
var matchingProperty = props.First(p => p.Name.Equals(codeProperty.Name);
var strLenAttr = (StringLengthAttribute) matchingPropery.GetCustomAttributes(typeof(StringLengthAttribute), true).FirstOrDefault();
if (strLenAttr == null) return 0;
return strLenAttr.MaximumLength;
}
return 0;
}
Related
OnModelCreating is called once per db context. This is a problem since the tenant Id is set per request.
How do I re-configure the global filter everytime I create an new instance of the dbcontext?
If I can't use global filter, what is the alternative way?
Update:
I needed to provide a generic filter with an expression like e => e.TenantId == _tenantId. I am using the following expression:
var p = Expression.Parameter(type, "e");
Expression.Lambda(
Expression.Equal(
Expression.Property(p, tenantIdProperty.PropertyInfo),
Expression.Constant(_tenantId))
p);
Since this is run once, _tenantId is fixed. So even if I update it, the first value is captured in the linq expression.
So my question, what is the proper way to set the right side of that equality.
With EF.Core you can actually use the following filter and syntax
protected void OnModelCreating(ModelBuilder modelBuilder)
{
var entityConfiguration = modelBuilder.Entity<MyTenantAwareEntity>();
entityConfiguration.ToTable("my_table")
.HasQueryFilter(e => EF.Property<string>(e, "TenantId") == _tenantProvider.GetTenant())
[...]
The _tenantProvider is the class responsible to get your tenant, in your case from the HttpRequest, to do it you can use HttpContextAccessor.
This is fixed with the following as the right expression
Expression.MakeMemberAccess(
Expression.Constant(this, baseDbContextType),
baseDbContextType.GetProperty("TenantId")
I use a base class for all my db contexts.
GetProperty() works as is because TenantId is public property.
and..
If you use it with soft delete, the solution is; --for ef core 3.1
internal static void AddQueryFilter<T>(this EntityTypeBuilder
entityTypeBuilder, Expression<Func<T, bool>> expression)
{
var parameterType = Expression.Parameter(entityTypeBuilder.Metadata.ClrType);
var expressionFilter = ReplacingExpressionVisitor.Replace(
expression.Parameters.Single(), parameterType, expression.Body);
var currentQueryFilter = entityTypeBuilder.Metadata.GetQueryFilter();
if (currentQueryFilter != null)
{
var currentExpressionFilter = ReplacingExpressionVisitor.Replace(
currentQueryFilter.Parameters.Single(), parameterType, currentQueryFilter.Body);
expressionFilter = Expression.AndAlso(currentExpressionFilter, expressionFilter);
}
var lambdaExpression = Expression.Lambda(expressionFilter, parameterType);
entityTypeBuilder.HasQueryFilter(lambdaExpression);
}
Usage:
if (typeof(ITrackSoftDelete).IsAssignableFrom(entityType.ClrType))
modelBuilder.Entity(entityType.ClrType).AddQueryFilter<ITrackSoftDelete>(e => IsSoftDeleteFilterEnabled == false || e.IsDeleted == false);
if (typeof(ITrackTenant).IsAssignableFrom(entityType.ClrType))
modelBuilder.Entity(entityType.ClrType).AddQueryFilter<ITrackTenant>(e => e.TenantId == MyTenantId);
Thanks to YZahringer
Is there a way that when I read an object from Realm that it can become a standalone or unmanaged object? In EF, this is called no tracking. The usage for this would be when I want to implement more business logic on my data objects before they are updated on the persistent data storage. I may want to give the RealmObject to a ViewModel, but when the changes come back from the ViewModel, I want to compare the disconnected object to the object in the datastore to determine what was changed, so If there was a way that I could disconnect the object from Realm when I give it to the ViewModel, then I can better manage what properties have changed, using my biz logic to do what I need, then save the changes back to realm.
I understand Realm does a lot of magic and many people will not want to add a layer like this but in my app, I cant really have the UI directly updating the datastore, unless there is a event that is raised that I can subscribe too and then attach my business logic this way.
I only saw one event and it does not appear to perform this action.
Thanks for your assistance.
First, get json NUGET :
PM> Install-Package Newtonsoft.Json
And, try this "hack" :
Deserialize the modified IsManaged property does the tricks.
public d DetachObject<d>(d Model) where d : RealmObject
{
return Newtonsoft.Json.JsonConvert.DeserializeObject<d>(
Newtonsoft.Json.JsonConvert.SerializeObject(Model)
.Replace(",\"IsManaged\":true", ",\"IsManaged\":false")
);
}
.
If you facing slow-down on JsonConvert:
According to source code
, the 'IsManaged' property only has get accessor and return true when private field _realm which is available
So, we has to set the instance of field _realm to null does the tricks
public d DetachObject<d>(d Model) where d : RealmObject
{
typeof(RealmObject).GetField("_realm",
System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.NonPublic)
.SetValue(Model, null);
return Model.IsManaged ? null : Model;
}
.
You will get empty RealmObject body after Realm are now implemented same strategy as LazyLoad
Record down live RealmObject and (deactivate) realm instance in object by Reflection. And set back recorded values to RealmObject. With handled all the ILists inside too.
public d DetachObject<d>(d Model) where d : RealmObject
{
return (d)DetachObjectInternal(Model);
}
private object DetachObjectInternal(object Model)
{
//Record down properties and fields on RealmObject
var Properties = Model.GetType().GetProperties(System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.Public)
.Where(x => x.Name != "ObjectSchema" && x.Name != "Realm" && x.Name != "IsValid" && x.Name != "IsManaged" && x.Name != "IsDefault")
.Select(x =>(x.PropertyType.Name == "IList`1")? ("-" + x.Name, x.GetValue(Model)) : (x.Name, x.GetValue(Model))).ToList();
var Fields = Model.GetType().GetFields(System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.Public)
.Select(x => (x.Name, x.GetValue(Model))).ToList();
//Unbind realm instance from object
typeof(RealmObject).GetField("_realm", System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.NonPublic).SetValue(Model, null);
//Set back the properties and fields into RealmObject
foreach (var field in Fields)
{
Model.GetType().GetField(field.Item1, System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.Public).SetValue(Model, field.Item2);
}
foreach (var property in Properties.OrderByDescending(x=>x.Item1[0]).ToList())
{
if (property.Item1[0] == '-')
{
int count = (int)property.Item2.GetType().GetMethod("get_Count").Invoke(property.Item2, null);
if (count > 0)
{
if (property.Item2.GetType().GenericTypeArguments[0].BaseType.Name == "RealmObject")
{
for (int i = 0; i < count; i++)
{
var seter = property.Item2.GetType().GetMethod("set_Item");
var geter = property.Item2.GetType().GetMethod("get_Item");
property.Item2.GetType().GetField("_realm", System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.NonPublic | System.Reflection.BindingFlags.Public).SetValue(property.Item2, null);
DetachObjectInternal(geter.Invoke(property.Item2, new object[] { i }));
}
}
}
}
else
{
Model.GetType().GetProperty(property.Item1, System.Reflection.BindingFlags.Instance | System.Reflection.BindingFlags.Public).SetValue(Model, property.Item2);
}
}
return Model;
}
.
For List of the RealmObject , Using Select():
DBs.All<MyRealmObject>().ToList().Select(t => DBs.DetachObject(t)).ToList();
.
(Java)You dont need this if youre in java:
Maybe some day, this feature will come to .NET Realm
Realm.copyFromRealm();
#xamarin #C# #Realm #RealmObject #detach #managed #IsManaged #copyFromRealm
Until its added to Realm for Xamarin, I added a property to my Model that creates a copy of the object. This seems to work for my use. The TwoWay Binding error messages are now also not an issue. For a more complicated application, I don't want to put business or data logic in the ViewModel. This allows all the Magic of xamarin forms to work and me to implement logic when its finally time to save the changes back to realm.
[Ignored]
public Contact ToStandalone()
{
return new Contact()
{
companyName = this.companyName,
dateAdded = this.dateAdded,
fullName = this.fullName,
gender = this.gender,
website = this.website
};
}
However, If there are any relationships this method does not work for the relationships. Copying the List is not really an option either as the relationship cant exist if the object is not attached to Realm, I read this some where, can't find it to ref now. So I guess we will be waiting for the additions to the framework.
Not currently in the Xamarin interface but we could add it. The Java interface already has copyFromRealm which performs a deep copy. That also has a paired merging copyToRealmOrUpdate.
See Realm github issue for further discussion.
However, as a design issue, is this really meeting your need in an optimal way?
I have used converters in WPF apps to insert logic into the binding - these are available in Xamarin Forms.
Another way in Xamarin forms is to use Behaviours, as introduced in the blog article and covered in the API.
These approaches are more about adding logic between the UI and ViewModel, which you could consider as part of the ViewModel, but before updates are propagated to bound values.
After wasting too much time in 3rd party libraries like AutoMapper, I've created my own extension function which works pretty well. This function simply uses Reflection with the recession. (Currently, Only for List type. You can extend the functionality for Dictionary and other types of the collection very easily or you can completely modify the functionality based on your own requirements.).
I didn't do too much time and complexity analysis. I've tested only for my test case which contains many nested RealmObject, built from a 3500+ line of JSON object, took only 15 milliseconds to clone the object.
Here the complete extension available via Github Gist. If you want to extend the functionality of this extension please update this Github Gist, So, other developers can take advantage of it.
Here the complete extension -
using System;
using System.Collections;
using System.Collections.Generic;
using System.Reflection;
using Realms;
namespace ProjectName.Core.Extensions
{
public static class RealmExtension
{
public static T Clone<T>(this T source) where T: new()
{
//If source is null return null
if (source == null)
return default(T);
var target = new T();
var targetType = typeof(T);
//List of skip namespaces
var skipNamespaces = new List<string>
{
typeof(Realm).Namespace
};
//Get the Namespace name of Generic Collection
var collectionNamespace = typeof(List<string>).Namespace;
//flags to get properties
var flags = BindingFlags.IgnoreCase | BindingFlags.DeclaredOnly | BindingFlags.Public | BindingFlags.Instance;
//Get target properties list which follows the flags
var targetProperties = targetType.GetProperties(flags);
//if traget properties is null then return default target
if (targetProperties == null)
return target;
//enumerate properties
foreach (var property in targetProperties)
{
//skip property if it's belongs to namespace available in skipNamespaces list
if (skipNamespaces.Contains(property.DeclaringType.Namespace))
continue;
//Get property information and check if we can write value in it
var propertyInfo = targetType.GetProperty(property.Name, flags);
if (propertyInfo == null || !property.CanWrite)
continue;
//Get value from the source
var sourceValue = property.GetValue(source);
//If property derived from the RealmObject then Clone that too
if (property.PropertyType.IsSubclassOf(typeof(RealmObject)) && (sourceValue is RealmObject))
{
var propertyType = property.PropertyType;
var convertedSourceValue = Convert.ChangeType(sourceValue, propertyType);
sourceValue = typeof(RealmExtension).GetMethod("Clone", BindingFlags.Static | BindingFlags.Public)
.MakeGenericMethod(propertyType).Invoke(convertedSourceValue, new[] { convertedSourceValue });
}
//Check if property belongs to the collection namespace and original value is not null
if (property.PropertyType.Namespace == collectionNamespace && sourceValue != null)
{
//get the type of the property (currently only supported List)
var listType = property.PropertyType;
//Create new instance of listType
var newList = (IList)Activator.CreateInstance(listType);
//Convert source value into the list type
var convertedSourceValue = Convert.ChangeType(sourceValue, listType) as IEnumerable;
//Enumerate source list and recursively call Clone method on each object
foreach (var item in convertedSourceValue)
{
var value = typeof(RealmExtension).GetMethod("Clone", BindingFlags.Static | BindingFlags.Public)
.MakeGenericMethod(item.GetType()).Invoke(item, new[] { item });
newList.Add(value);
}
//update source value
sourceValue = newList;
}
//set updated original value into the target
propertyInfo.SetValue(target, sourceValue);
}
return target;
}
}
}
I'm currently working in a project that involves the encryption of a few columns in an existing database. There is quite a lot of code already written against the current schema, a lot of which is in the form of custom linq-to-sql queries. The number of queries is in the neighbourhood of a 5 figure number, so modifying and re-testing each and everyone of them would be way too expensive.
An alternative we found is to keep the DB schema the same --only altering the columns length slightly, which mean we don't need to change our current entity class definitions-- and instead, changing the expression trees on-the-fly, before they reach the l2sql IQueryProvider, and apply a decryption function on the columns I need. I do this by wrapping the pertinent Table<TEntity> properties of my DataContext with a custom IQueryable<TEntity> implementation, which allows me to preview every single query in the system.
In my current implementation, say I've got this query:
var mydate = new DateTime(2013, 1, 1);
var context = new DataContextFactory.GetClientsContext();
Expression<Func<string>> foo = context.MyClients.First(
c => c.BirthDay < mydate).EncryptedColumn;
but when I catch the query, I change it to read:
Expression<Func<string>> foo = context.Decrypt(
context.MyClients.First(c => c.BirthDay < mydate).EncryptedColumn);
I do this using the ExpressionVisitor class. In the VisitMember method, I check and see whether the current MemberExpression refers to an encrypted column. If it does, I substitute the expression for a method call:
private const string FuncName = "Decrypt";
protected override Expression VisitMember(MemberExpression ma)
{
if (datactx != null && IsEncryptedColumnReference(ma))
return MakeCallExpression(ma);
}
return base.VisitMember(ma);
}
private static bool IsEncryptedColumnReference(MemberExpression ma)
{
return ma.Member.Name == "EncryptedColumn"
&& ma.Member.DeclaringType == typeof(MyClient);
}
private Expression MakeCallExpression(MemberExpression ma)
{
const BindingFlags flags = BindingFlags.Instance | BindingFlags.Public;
var mi = typeof(MyDataContext).GetMethod(FuncName, flags);
return Expression.Call(datactx, mi, ma);
}
datactx is an instance variable with a reference to the expression pointing at the current datacontext (which I look up in a previous pass).
My problem is that if I have a query such as:
var qbeClient = new MyClient { EncryptedColumn = "FooBar" };
Expression<Func<MyClient>> dbquery = () => context.MyClients.First(
c => c.EncryptedColumn == qbeClient.EncryptedColumn);
I want it to be turned into:
Expression<Func<MyClient>> dbquery = () => context.MyClients.First(c =>
context.Decrypt(c.EncryptedColumn) == qbeClient.EncryptedColumn);
instead, what I'm getting is this:
Expression<Func<MyClient>> dbquery = () => context.MyClients.First(c =>
context.Decrypt(c.EncryptedColumn) == context.Decrypt(qbeClient.EncryptedColumn));
Which I don't want, because when I've got an in-memory object, the data is already unencrypted (besides, I don't want a nasty db function call against my objects!)
So, that's basically my question: Having a MemberExpression instance, how can I determine whether it refers to an in-memory object or a row in the database?
Thanks in advance
Edit:
#Shlomo's code actually solves the case I posted, but now one of my previous tests got broken:
var context = new DataContextFactory.GetClientsContext();
Expression<Func<string>> expr = context.MyClients.First().EncryptedColumn;
Expression<Func<string>> expected = context.Decrypt(
context.MyClients.First().EncryptedColumn);
var actual = MyVisitor.Visit(expr);
Assert.AreEqual(expected.ToString(), actual.ToString());
In this case, the reference to EncryptedColumn isn't a parameter, but it should definitely be taken into account by the visitor!
A MemberExpression representing a DB row will be a descendent of a ParameterExpression. In-Memory objects will not, they'll most likely come from some form of a FieldExpression.
In your case, something like this will work for most cases (adding one method to your code, and revising your VisitMember method:
private bool IsFromParameter(MemberExpression ma)
{
if(ma.Expression.NodeType == ExpressionType.Parameter)
return true;
if(ma.Expression is MemberExpression)
return IsFromParameter(ma.Expression as MemberExpression);
return false;
}
protected override Expression VisitMember(MemberExpression ma)
{
if (datactx != null && IsEncryptedColumnReference(ma) && IsFromParameter(ma))
return MakeCallExpression(ma);
}
return base.VisitMember(ma);
}
I am having some trouble with a linq query I am trying to write.
I am trying to use the repository pattern without to much luck. Basically I have a list of transactions and a 2nd list which contains the description field that maps against a field in my case StoreItemID
public static IList<TransactionViewModel> All()
{
var result = (IList<TransactionViewModel>)HttpContext.Current.Session["Transactions"];
if (result == null)
{
var rewardTypes = BusinessItemRepository.GetItemTypes(StoreID);
HttpContext.Current.Session["Transactions"] =
result =
(from item in new MyEntities().TransactionEntries
select new TransactionViewModel()
{
ItemDescription = itemTypes.FirstOrDefault(r=>r.StoreItemID==item.StoreItemID).ItemDescription,
TransactionDate = item.PurchaseDate.Value,
TransactionAmount = item.TransactionAmount.Value,
}).ToList();
}
return result;
}
public static List<BusinessItemViewModel>GetItemTypes(int storeID)
{
var result = (List<BusinessItemViewModel>)HttpContext.Current.Session["ItemTypes"];
if (result == null)
{
HttpContext.Current.Session["ItemTypes"] = result =
(from items in new MyEntities().StoreItems
where items.IsDeleted == false && items.StoreID == storeID
select new BusinessItemViewModel()
{
ItemDescription = items.Description,
StoreID = items.StoreID,
StoreItemID = items.StoreItemID
}).ToList();
}
return result;
However I get this error
Unable to create a constant value of type 'MyMVC.ViewModels.BusinessItemViewModel'. Only primitive types ('such as Int32, String, and Guid') are supported in this context.
I know its this line of code as if I comment it out it works ok
ItemDescription = itemTypes.FirstOrDefault(r=>r.StoreItemID==item.StoreItemID).ItemDescription,
How can I map ItemDescription against my list of itemTypes?
Any help would be great :)
This line has a problem:
ItemDescription = itemTypes.FirstOrDefault(r=>r.StoreItemID==item.StoreItemID)
.ItemDescription,
Since you are using FirstOrDefault you will get null as default value for a reference type if there is no item that satifies the condition, then you'd get an exception when trying to access ItemDescription - either use First() if there always will be at least one match or check and define a default property value for ItemDescription to use if there is none:
ItemDescription = itemTypes.Any(r=>r.StoreItemID==item.StoreItemID)
? itemTypes.First(r=>r.StoreItemID==item.StoreItemID)
.ItemDescription
: "My Default",
If itemTypes is IEnumerable then it can't be used in your query (which is what the error message is telling you), because the query provider doesn't know what to do with it. So assuming the that itemTypes is based on a table in the same db as TransactionEntities, then you can use a join to achieve the same goal:
using (var entities = new MyEntities())
{
HttpContext.Current.Session["Transactions"] = result =
(from item in new entities.TransactionEntries
join itemType in entities.ItemTypes on item.StoreItemID equals itemType.StoreItemID
select new TransactionViewModel()
{
ItemDescription = itemType.ItemDescription,
TransactionDate = item.PurchaseDate.Value,
TransactionAmount = item.TransactionAmount.Value,
CustomerName = rewards.CardID//TODO: Get customer name
}).ToList();
}
I don't know the structure of your database, but hopefully you get the idea.
I had this error due a nullable integer in my LINQ query.
Adding a check within my query it solved my problem.
query with problem:
var x = entities.MyObjects.FirstOrDefault(s => s.Obj_Id.Equals(y.OBJ_ID));
query with problem solved:
var x = entities.MyObjects.FirstOrDefault(s => s.Obj_Id.HasValue && s.Obj_Id.Value.Equals(y.OBJ_ID));
I have a property in umbraco that uses a drop down data type with a set of prevalues that you can select from.
How do I retrieve a list of all the possible prevalues that are in this drop down list?
There's a helper method in umbraco.library that does that.
From xslt:
<xsl:variable name="prevalues" select="umbraco.library:GetPreValues(1234)" />
From code:
using umbraco;
XPathNodeIterator prevalues = library.GetPrevalues(1234);
Replace 1234 with the id of your datatype (You can see it in the bottom of your browser when hovering your mouse over the datatype in the developers section)
Regards
Jesper Hauge
Here is the code that I use in one of my Umbraco datatypes to get a DropDownList containing all possible prevalues:
var prevalues = PreValues.GetPreValues(dataTypeDefinitionId);
DropDownList ddl = new DropDownList();
if (prevalues.Count > 0)
{
for (int i = 0; i < prevalues.Count; i++)
{
var prevalue = (PreValue)prevalues[i];
if (!String.IsNullOrEmpty(prevalue.Value))
{
ddl.Items.Add(new ListItem(prevalue.Value, prevalue.DataTypeId.ToString()));
}
}
}
Replace dataTypeDefinitionId with the id of your datatype.
I know this is an old question, but I created this method based on the information provided in this answer and I think it is worth documenting:
public static class UmbracoExtensions
{
public static IEnumerable<string> GetDropDownDataTypeValues(int dataTypeId)
{
var dataTypeValues = umbraco.library.GetPreValues(dataTypeId);
var dataTypeValuesEnumerator = dataTypeValues.GetEnumerator();
while (dataTypeValues.MoveNext())
{
dynamic dataTypeItem = dataTypeValues.Current;
yield return dataTypeItem.Value;
}
}
}