I have a query which uses code like:
criteria.AddOrder(
Order.Asc(
Projections.SqlFunction(
new StandardSQLFunction("NEWID"),
new NHType.GuidType(),
new IProjection[0])));
The purpose is to get a random ordering. I run this against an SQL server, but I would also like to run it against an SQLite, since we use this for testing. SQLite does not support NEWID() but has Random instead. Is it possible to write the code (or configure) such that the same query will work against both databases?
I think the way to do this is to create two custom dialects. Have each one implement a random function differently:
public class MyMsSqlDialect : MsSql2012Dialect
{
protected override void RegisterFunctions()
{
base.RegisterFunctions();
RegisterFunction("random", new StandardSQLFunction("NEWID", NHibernateUtil.Guid));
}
}
public class MySqliteDialect : SQLiteDialect
{
protected override void RegisterFunctions()
{
base.RegisterFunctions();
RegisterFunction("random", new StandardSQLFunction("random", NHibernateUtil.Guid));
}
}
Now, the following query should work fine in either database:
criteria.AddOrder(
Order.Asc(
Projections.SqlFunction("random", NHibernateUtil.Guid)));
Note that this is cheating a bit. random doesn't return a Guid in the Sqlite flavor, but since NHibernate doesn't need that information to do the ORDER BY, nothing should go wrong. I would consider calling it random_order or something to make it clear that this is only for ordering.
Related
To make updates to a record of SQL Server using Entity Framework Core, I query the record I need to update, make changes to the object and then call .SaveChanges(). This works nice and clean.
For example:
var emp = _context.Employee.FirstOrDefault(item => item.IdEmployee == Data.IdEmployee);
emp.IdPosition = Data.IdPosition;
await _context.SaveChangesAsync();
But is there a standard method if I want to update multiple records?
My first approach was using a list passing it to the controller, but then I would need to go through that list and save changes every time, never really finished this option as I regarded it as not optimal.
For now what I do is instead of passing a list to the controller, I pass each object to the controller using a for. (kind of the same...)
for(int i = 0; i < ObjectList.Count; i ++)
{
/* Some code */
var httpResponseObject = await MyRepositories.Post<Object>(url+"/Controller", Object);
}
And then do the same thing on the controller as before, when updating only one record, for each of the records...
I don't feel this is the best possible approach, but I haven't found another way, yet.
What would be the optimal way of doing this?
Your question has nothing to do with Blazor... However, I'm not sure I understand what is the issue. When you call the SaveChangesAsync method, all changes in your context are committed to the database. You don't have to pass one object at a time...You can pass a list of objects
Hope this helps...
Updating records in bulk using Entity Framework or other Object Relational Mapping (ORM) libraries is a common challenge because they will run an UPDATE command for every record. You could try using Entity Framework Plus, which is an extension to do bulk updates.
If updating multiple records with a single call is critical for you, I would recommend just writing a stored procedure and call if from your service. Entity Framework can also run direct queries and stored procedures.
It looks like the user makes some changes and then a save action needs to persist multiple records at the same time. You could trigger multiple AJAX calls—or, if you need, just one.
What I would do is create an endpoint—with an API controller and an action—that's specific to your needs. For example, to update the position of records in a table:
Controller:
/DataOrder
Action:
[HttpPut]
public async void Update([FromBody] DataChanges changes)
{
foreach(var change in changes)
{
var dbRecord = _context.Employees.Find(change.RecordId);
dbRecord.IdPosition = change.Position;
}
_context.SaveChanges();
}
public class DataChanges
{
public List<DataChange> Items {get;set;}
public DataChangesWrapper()
{
Items = new List<DataChange>();
}
}
public class DataChange
{
public int RecordId {get;set;}
public int Position {get;set;}
}
The foreach statement will execute an UPDATE for every record. If you want a single database call, however, you can write a SQL query or have a stored procedure in the database and pass the data as a DataTable parameter instead.
I have used Partitioned scheme and used data-affinity to store related data together on Cluster member.
I am only interested to execute my Invocable on a subset of entire data. Can I find specific members storing subset of data I am interested in?
Currently I do this: partitionedService.getOwnershipEnabledMembers() which gets me all nodes running this service in cluster.
You can use PartitionedService.getKeyOwner(Object key)
You can use an EntryProcessor and execute it against the data that you want to process, eg.
public class MyProc extends AbstractProcessor implements Serializable
{
public Object process(Entry entry)
{
System.out.println("*** processing " + entry.getKey());
}
}
Now to run that, just:
cache.invokeAll(collKeys, new MyProc());
... where "collKeys" is a collection of keys to your cache, or:
cache.invokeAll(filter, new MyProc());
... where "filter" is a query on your cache.
I have a simple page with a Grid that I'm binding a collection of objects to. I also have some simple functionality on the grid to edit and save rows. I would like to write unit tests for this page, but it's not really making sense to me.
For example:
Private Sub LoadGrid()
'Populate Collection
grid.datasource = MyCollection
grid.databind()
end sub
I guess a Sub really doesn't need a unit test, but what if this were a function that returned true when the grid had been loaded. How do you write a unit test for this? What other test should be done on a simple web page like this?
As always, thanks to everyone who give any sort of input.
How do you write a unit test for this?
The first step is actually making your form testable. Have a look at this page for separating UI and BL layers, there are about a bajillion different ways to implement MVC, MVP, and all of its variants, and there's no One True Way™ to do it. So long as your code is sane and consistent, other people will be able to work on your code.
I personally find the following pattern works in most cases for testing UIs:
Create an interface representing your Model.
Create a class for your controller which handles all the updates to the model.
Your view should listen to changes to the model.
So in the end, you end up with something like this (sorry, my VB-fu is rusty, writing this in C# instead):
interface IProductPageModel
{
int CurrentPage { get; set; }
int ItemsPerPage { get; set; }
DataSet ProductDataSet { get; set; }
}
class ProductPageController
{
public readonly IProductPageModel Model;
public ProductPageController(IProductPageModel model)
{
this.Model = model;
}
public void NavigateTo(int page)
{
if (page <= 0)
throw new ArgumentOutOfRangeException("page should be greater than 0");
Model.CurrentPage = page;
Model.ProductDataSet = // some call to retrieve next page of data
}
// ...
}
This is concept code, of course, but you can see how its very easy to unit test. In principle, you could re-use the same controller code in for desktop apps, silverlight, etc since your controller doesn't depend directly on any particular view implementation.
And finally on your form side, you'd implement your page similar to:
public class ProductPage : Page, IProductPageModel
{
ProductPageController controller;
public ProductPage()
{
controller = new ProductPageController(this);
}
public DataSet ProductDataSet
{
get { return (DataSet)myGrid.DataSource; }
set { myGrid.DataSource = value; myGrid.DataBind(); }
}
protected void NavigateButton_OnCommand(object sender, CommandEventArgs e)
{
controller.NavigateTo(Convert.ToInt32(e.CommandArgument));
}
}
Here there's no real distinction between view and model -- they're the same entity. The idea is to make your code-behind as "stupid" as possible, so that as much testable business logic is contained in the controller as possible.
What other test should be done on a
simple webpage like this?
You'd want tests for any sort of form validation, you want to make sure you're throwing exceptions in exceptional cases, ensuring that your controller methods update your model in an expected way, and so on.
Juliet is right.
The line of code where you said
'Populate Collection
that is the testable part. You can do assertions on if the collection is null, if it has items, if it has exactly 42 items. But that would be an integration test.
If you can isolate all the calls to the database (the part that returns a datareader), then return a empty, fake DbDataReader, then you can test everything inbetween the UI and the database.
Tests that spin up a browser and verify that a table is render, similarly is a integration test that will depend on having IIS up and working (and a DB up and working, unless you have a repository you can fake)
If you are just getting started, I would look for all the easy to test code first, such as methods that do have dependencies on the database, then move on the tricker tests that require mocking/stubbing/faking database servers, etc.
Using FluentMigrator, the default creation of a Column using .AsString() results in an nvarchar(255). Is there a simple way (before I modify the FluentMigrator code) to create a column of type nvarchar(MAX)?
You could create an extension method to wrap .AsString(Int32.MaxValue) within .AsMaxString()
e.g.
internal static class MigratorExtensions
{
public static ICreateTableColumnOptionOrWithColumnSyntax AsMaxString(this ICreateTableColumnAsTypeSyntax createTableColumnAsTypeSyntax)
{
return createTableColumnAsTypeSyntax.AsString(int.MaxValue);
}
}
OK, I found it. Basically, use .AsString(Int32.MaxValue). Pity there's not a .AsMaxString() method, but I guess it's easy enough to put in...
You can use AsCustom("nvarchar(max)") and pack it to extension
If you often create columns/tables with the same settings or groups of columns, you should be creating extension methods for your migrations!
For example, nearly every one of my tables has CreatedAt and UpdatedAt DateTime columns, so I whipped up a little extension method so I can say:
Create.Table("Foos").
WithColumn("a").
WithTimestamps();
I think I created the Extension method properly ... I know it works, but FluentMigrator has a LOT of interfaces ... here it is:
public static class MigrationExtensions {
public static ICreateTableWithColumnSyntax WithTimestamps(this ICreateTableWithColumnSyntax root) {
return root.
WithColumn("CreatedAt").AsDateTime().NotNullable().
WithColumn("UpdatedAt").AsDateTime().NotNullable();
}
}
Similarly, nearly every one of my tables has an int primary key called 'Id', so I think I'm going to add Table.CreateWithId("Foos") to always add that Id for me. Not sure ... I actually just started using FluentMigrator today, but you should always be refactoring when possible!
NOTE: If you do make helper/extension methods for your migrations, you should never ever ever change what those methods do. If you do, someone could try running your migrations and things could explode because the helper methods you used to create Migration #1 works differently now than they did earlier.
Here is the code for creating columns incase it helps you create helper methods: https://github.com/schambers/fluentmigrator/blob/master/src/FluentMigrator/Builders/Create/Column/CreateColumnExpressionBuilder.cs
How about extending like this:
public static class StringMaxMigratorExtensions
{
public static ICreateTableColumnOptionOrWithColumnSyntax AsStringMax(this ICreateTableColumnAsTypeSyntax createTableColumnAsTypeSyntax)
{
return createTableColumnAsTypeSyntax.AsCustom("nvarchar(max)");
}
public static IAlterColumnOptionSyntax AsStringMax(this IAlterColumnAsTypeSyntax alterColumnAsTypeSyntax)
{
return alterColumnAsTypeSyntax.AsCustom("nvarchar(max)");
}
}
I'm using Python+PyAMF to talk back and forth with Flex clients, but I've run into a problem with the psudo-Enum-Singletons I'm using:
class Type {
public static const EMPTY:Type = new Type("empty");
public static const FULL:Type = new Type("full");
...
}
When I'm using locally created instances, everything is peachy:
if (someInstance.type == Type.EMPTY) { /* do things */ }
But, if 'someInstance' has come from the Python code, it's instance of 'type' obviously won't be either Type.EMPTY or Type.FULL.
So, what's the best way to make my code work?
Is there some way I can control AMF's deserialization, so when it loads a remote Type, the correct transformation will be called? Or should I just bite the bullet and compare Types using something other than ==? Or could I somehow trick the == type cohesion into doing what I want?
Edit: Alternately, does Flex's remoting suite provide any hooks which run after an instance has been deserialized, so I could perform a conversion then?
Random thought: Maybe you could create a member function on Type that will return the canonical version that matches it?
Something like:
class Type {
public static const EMPTY:Type = new Type("empty");
public static const FULL:Type = new Type("full");
...
// I'm assuming this is where that string passed
// in to the constructor goes, and that it's unique.
private var _typeName:String;
public function get canonical():Type {
switch(this._typeName) {
case "empty": return EMPTY;
case "full": return FULL;
/*...*/
}
}
}
As long as you know which values come from python you would just convert them initially:
var fromPython:Type = /*...*/
var t:Type = fromPython.canonical;
then use t after that.
If you can't tell when things come from python and when they're from AS3 then it would get pretty messy, but if you have an isolation layer between the AS and python code you could just make sure you do the conversion there.
It's not as clean as if you could control the deserialization, but as long as you've got a good isolation layer it should work.