Friday, July 27, 2012

Dealing with DataSet’s Outdated Nullability Model

The DataSet is one of the most ancient relics of the .NET framework. DataSets were designed before the invention of Nullable<T> and came up with a different way to deal with database null values. So we are stuck with properties that throw StrongTypingException when the property is null and special IsXXXNull() methods to test whether a property is null. So you end up with ugly code like this:

if (!row.IsReportsToNull())
{
Console.WriteLine(row.ReportsTo);
}
else
{
Console.WriteLine("NULL");
}

or

try
{
Console.WriteLine(row.ReportsTo);
}
catch (StrongTypingException)
{
Console.WriteLine("NULL");
}

This code dance is necessary because you can’t even touch a null property without triggering an exception. So you can’t pass the property to a method that would check for nullability. But you can pass a Func<T> to a lamda:

private string GetStringForTFromDataRowField<T>(Func<bool> isNull, Func<T> getValue, 
string formatString = "{0}")
{
if (!isNull())
{
return string.Format(formatString, getValue());
}
else
{
return "NULL";
}
}

With this function I can reference the field in one (ugly) statement:

Console.WriteLine(GetStringForTFromDataRowField(row.IsReportsToNull, 
() => { return row.ReportsTo; }));

Here I am not referencing the actual property until I confirm that the property isn’t null. Since the property is referenced in an anonymous method it isn’t evaluated until after the IsNull() method confirms that the value isn’t null.

What do I get? Slightly cleaner code, I guess. You still need to match the property with the correct IsXXXNull() method. I find the code in the first couple of examples so ugly that this hack is worth the effort for me.

Thursday, June 14, 2012

Using LINQPad’s SQL Tab To Investigate EF’s Crud Commands

I like LINQPad. One thing I like to do with it is to use to figure out what queries the Entity Framework is generating when I do basic tasks. In this post, I will look at what is going on for basic CRUD operations (actually CUD, since the Read operations are self evident). For this example, I am going use the method outlined in Querying EF ObjectContext from LINQPad; I am using the PUBS database and the model from Building an EF 4 Model Assembly for Pubs DB.

Whenever you run a query in LINQPad you can click on the SQL to see what has been sent to the server. I find this method more useful than Getting TSQL Query With ToTraceString() in EF; I can run code that is not a LINQ query and see what happens.

sql_tab 

Add

Given this function:

void addTitle(PubsModel.pubsEntities context)
{
     var newTitle = new title();
     newTitle.title_id = "myTtl";
     newTitle.title = "Hamlet";
     newTitle.type = "Literature";
     newTitle.pubdate = DateTime.Today;
     newTitle.publisher = 
          context.publishers.Where(p => p.pub_id == "9901").Single();

     context.titles.AddObject(newTitle);

     context.SaveChanges();
}

I get this SQL from LINQPad:

SELECT TOP (2) 
[Extent1].[pub_id] AS [pub_id], 
[Extent1].[pub_name] AS [pub_name], 
[Extent1].[city] AS [city], 
[Extent1].[state] AS [state], 
[Extent1].[country] AS [country]
FROM [dbo].[publishers] AS [Extent1]
WHERE '9901' = [Extent1].[pub_id]
GO

-- Region Parameters
DECLARE @0 VarChar(6) = 'myTtl'
DECLARE @1 VarChar(80) = 'Hamlet'
DECLARE @2 Char(12) = 'Literature'
DECLARE @3 Char(4) = '9901'
DECLARE @4 DateTime2 = '2012-04-07 00:00:00.0000000'
-- EndRegion
insert [dbo].[titles]([title_id], [title], [type], [pub_id], [price], [advance], [royalty], [ytd_sales], [notes], [pubdate])
values (@0, @1, @2, @3, null, null, null, null, null, @4)


In the top part, EF is getting the linked record from the publishers table. For this specific add, the get is probably not necessary. I could avoid that call setting the pub_id directly as opposed to setting the publisher navigation property to the row.

The second part is a basic INSERT query with parameters. Other than the cryptic names and explicitly referencing null in the VALUES clause, this is pretty close to what I would have written.

Edit


Given this function:

void editTitle(PubsModel.pubsEntities context)
{
     var myTitle = context.titles.Where(t => t.title_id == "myTtl").Single();

     myTitle.price = 100m;
     myTitle.advance = 1000000m;
    
     context.SaveChanges();
}

I get this SQL from LINQPad:

SELECT TOP (2) 
[Extent1].[title_id] AS [title_id], 
[Extent1].[title] AS [title], 
[Extent1].[type] AS [type], 
[Extent1].[pub_id] AS [pub_id], 
[Extent1].[price] AS [price], 
[Extent1].[advance] AS [advance], 
[Extent1].[royalty] AS [royalty], 
[Extent1].[ytd_sales] AS [ytd_sales], 
[Extent1].[notes] AS [notes], 
[Extent1].[pubdate] AS [pubdate]
FROM [dbo].[titles] AS [Extent1]
WHERE 'myTtl' = [Extent1].[title_id]
GO

-- Region Parameters
DECLARE @0 Decimal(7,4) = 100
DECLARE @1 Decimal(11,4) = 1000000
DECLARE @2 VarChar(6) = 'myTtl'
-- EndRegion
update [dbo].[titles]
set [price] = @0, [advance] = @1
where ([title_id] = @2)


Not much to see here. It is getting the data with a SELECT command and then editing it with the data with an UPDATE command. Again this is pretty close to what I would do.


Delete


Given this function:

void delTitle(PubsModel.pubsEntities context)
{
     var myTitle = context.titles.Where (t => t.title_id == "myTtl").Single ();
     context.DeleteObject(myTitle);
     context.SaveChanges();
}

I get this SQL from LINQPad:

SELECT TOP (2) 
[Extent1].[title_id] AS [title_id], 
[Extent1].[title] AS [title], 
[Extent1].[type] AS [type], 
[Extent1].[pub_id] AS [pub_id], 
[Extent1].[price] AS [price], 
[Extent1].[advance] AS [advance], 
[Extent1].[royalty] AS [royalty], 
[Extent1].[ytd_sales] AS [ytd_sales], 
[Extent1].[notes] AS [notes], 
[Extent1].[pubdate] AS [pubdate]
FROM [dbo].[titles] AS [Extent1]
WHERE 'myTtl' = [Extent1].[title_id]
GO

-- Region Parameters
DECLARE @0 VarChar(6) = 'myTtl'
-- EndRegion
delete [dbo].[titles]
where ([title_id] = @0)

This one is a little bit more controversial. Do I really need to get the doomed record before I delete it? Does it really make that much difference? I probably wouldn’t get the record before I deleted it. I have seen examples where you send an update command through EF that deletes the record without ever reading it.


LINQPad Code

const string EF_CONNECTION_STRING = 
@"metadata=res://*/PubsModel.csdl|res://*/PubsModel.ssdl|res://*/PubsModel.msl;" +
@"provider=System.Data.SqlClient;" +
@"provider connection string=" +
@"'data source=WIN-SEVEN-02\SQLEXPRESS;initial catalog=pubs;" +
@"integrated security=True;" + 
@"MultipleActiveResultSets=True;App=EntityFramework'";
void Main()
{
PubsModel.pubsEntities context = new PubsModel.pubsEntities(EF_CONNECTION_STRING);
addTitle(context);
showMyTitle(context, "After Add");
editTitle(context);
showMyTitle(context, "After Edit");
delTitle(context);
showMyTitle(context, "After Delete");

}
void addTitle(PubsModel.pubsEntities context)
{
var newTitle = new title
{
title_id = "myTtl",
title1 = "Hamlet",
type = "Literature",
pubdate = DateTime.Today,
publisher = context.publishers.Where(p => p.pub_id == "9901").Single()
// If I set the ID, I can avoid a SELECT command
//pub_id = "9901"
};
context.titles.AddObject(newTitle);
context.SaveChanges();
}
void editTitle(PubsModel.pubsEntities context)
{
var myTitle = context.titles.Where(t => t.title_id == "myTtl").Single();

//    myTitle.EntityState.Dump("EntityState Before Edit");
isDirty(myTitle).Dump("isDirty Before Edit");
myTitle.price = 100m;
myTitle.advance = 1000000m;
//    myTitle.EntityState.Dump("EntityState After Edit");
isDirty(myTitle).Dump("isDirty After Edit");

context.SaveChanges();
}
void delTitle(PubsModel.pubsEntities context)
{
var myTitle = context.titles.Where (t => t.title_id == "myTtl").Single ();
context.DeleteObject(myTitle);
context.SaveChanges();
}
void showMyTitle(PubsModel.pubsEntities context, string message)
{
context.titles.Include("publisher").Where(t => t.title_id == "myTtl").Dump(message);
}
bool isDirty(EntityObject entity)
{
return entity.EntityState == EntityState.Added || 
entity.EntityState == EntityState.Modified || 
entity.EntityState == EntityState.Deleted;
}

Thursday, June 07, 2012

Improv @ Seattle Code Camp

Ted Neward is doing a session on Improv at the 2012 Seattle Code Camp. So I am excited to go to a code camp to take a tiny Improv workshop. As I said here, I have been doing improve for the past couple of years. My worlds come together.

I have been doing a lot of thinking about how Improv would benefit many of us nerds and geeks (Geekprov anyone?) You know the classic stereotype of a computer programmer: a socially awkward recluse who spends whole nights in a dark room coding with a 2 liter of Mountain Dew and cold pizza. I am the poster child.

Since I have started to improvise, I have been braver; I have spoken in public several times. I feel that I listen better and feel better equipped to deal with change. I can better deal with fear and failure. When I find myself in a Dilbertian situation, I think about how I would use it in a scene. I believe other geeks could benefit from Improv.

I have toyed with the idea of bring Improv to the unwashed developer masses. I’m not sure how I would do it in an hour session. Would it be lecture or practice? What would I do for the hour? I am looking forward to seeing what Ted has in mind. Ted Neward is Cool!

Thursday, May 31, 2012

On Not Writting

For the past month I have struggled to finish what I've started. I think too much about what I'm trying to say and nothing comes out. Or my inner critic edits it before I can finish the bloody sentence. It can always be better.

It once was better: I start thinking about something, say on the way home from work, walking through Riverfront Park (doing laundry, driving to nowhere, etc). Regardless of what I was doing, it was brilliant. The kind of thing that changes the world. Everyone would recognize ME as the guru of whatever. So on and you know the rest… Then I get home, open up Word and… The phrases that sounded good in my head look terrible on the screen (if I can even remember them).

As the Improv Gods would say, I am in my head. I am thinking too much about things that are not important, at least not in the present moment. If I can get the words down, I can fix the style later. If I can't get the words down in the first place I have nothing. Ok, Mr. Critic it ain't perfect but it's on virtual paper.

Ok, I let Mr. Critic silence me. The only way I can grow is to put myself out there and risk being the fool. I can’t succeed if I don’t risk failure. This post, like the others, is a disappointment. It had so much promise when I was thinking it over in my mind. .. It may not meet my impossible, I'll post it anyway.

Saturday, May 12, 2012

Log4Net: Minimalist demo using config settings

The other day, it was decided at work that we will be using Log4Net, Apache’s logging tool for the .NET Framwork. It was left to me to investigate this tool. This post is the beginning of this investigation. For my first post, I present a minimal demo of Log4Net using Config settings. So here are the steps I’ve come up with using Microsoft C# Express 2010:

  1. Create a Console Application Project
  2. Make sure that Target Framework is .NET Framework 4, (select the project in Solution Explorer and press Alt+Enter, Look at Target Framework), the default is .NET Framework 4 Client Profile.
  3. Add a reference to log4net.dll
  4. Add the following attribute outside of any class or namespace (in my project I put it in Program.cs between the using and namespace statements, in a real project I would probably put it in AssemblyInfo.cs):
    [assembly: log4net.Config.XmlConfigurator]
    
    In this case XmlConfigurator tells the Framework to look in my application's .config file for configuration information.
  5. In app.config:

    <configSections> <section name="log4net" type="log4net.Config.Log4NetConfigurationSectionHandler, log4net"/> </configSections>

    This tells the framework about Log4Net’s custom configuration section below.

    <log4net> <appender name="MyAppender" type="log4net.Appender.FileAppender"> <file value="TestLog.txt"/> <layout type="log4net.Layout.SimpleLayout" /> </appender>


    log4net is the configuration section we described configSections above.


    This creates an appender; an appender is an object that is writes the logged data to somewhere. A FileAppender will write it to a text file. The name is used below to tell log4net to actually use this appender;


    The file element declares the name of the file. Here we will be logging to TestLog.txt in the same directory as the program.


    The layout tells log4net how to write the data. SimpleLayout is as simple as they get, In the real world, I’d use a more flexible layout. the PatternLayout is popular.

    <root>
    <appender-ref ref="MyAppender"/>
    </root>
    </log4net>

    Declaring an appender is good, but log4net won’t use it unless you tell it to.  The appender-ref element tells log4net to use my appender. You can specify as many appenders as you want.


  6. Create an instance of Log
    log4net.ILog log = log4net.LogManager.GetLogger("Some String");
    

  7. And Log away.
    log.Info("For Your Information");
    log.Warn("Danger Will Robinson");
    log.Debug("Debug Message");
    try
    {
    throw new Exception("Deliberate Exception");
    }
    catch (Exception ex)
    {
    log.Error(ex.Message, ex);
    }
    log.Fatal("Die. Die. Die!", new ApplicationException("Fatal Exception"));

    Running the code above, my log looks like this (note that there are not time stamps):

    INFO - For Your Information
    WARN - Danger Will Robinson
    DEBUG - Debug Message
    ERROR - Deliberate Exception
    System.Exception: Deliberate Exception
    at Log4NetConsole001.Program.Main(String[] args) in C:\Users\jack\documents\visual studio 2010\Projects\Log4NetConsole001\Log4NetConsole001\Program.cs:line 20
    FATAL - Die. Die. Die!
    System.ApplicationException: Fatal Exception
    
    


 

That is my minimalist implementation of Log4Net; it is probably too simple to be useful, but is is a start. If my ambition holds, I hope to expand on this.

Complete Sample

program.cs

using System;

[assembly: log4net.Config.XmlConfigurator]

namespace Log4NetConsole001
{
class Program
{
static void Main(string[] args)
{
log4net.ILog log = log4net.LogManager.GetLogger("Some String");
log.Info("For Your Information");
log.Warn("Danger Will Robinson");
log.Debug("Debug Message");
try
{
throw new Exception("Deliberate Exception");
}
catch (Exception ex)
{
log.Error(ex.Message, ex);
}
log.Fatal("Die. Die. Die!", new ApplicationException("Fatal Exception"));
Console.WriteLine("======================= End ========================");
Console.ReadKey();
}
}
}

app.config

<?xml version="1.0"?>
<configuration>
<configSections>
<section name="log4net" 
     type="log4net.Config.Log4NetConfigurationSectionHandler, log4net"/>
</configSections>

<log4net>
  <appender name="MyAppender" type="log4net.Appender.FileAppender">
<file value="TestLog.txt"/>
<layout type="log4net.Layout.SimpleLayout" />
  </appender>

<root>
<appender-ref ref="MyAppender"/>
</root>
</log4net>
<startup>
<supportedRuntime version="v4.0" sku=".NETFramework,Version=v4.0"/>
</startup>
</configuration>

Thursday, April 26, 2012

My Nokia 900

Recently I traded my iPhone 3G for a Nokia 900. The Nokia is nicer than the Apple, but I don't think that is a fair comparison; lots have happened in the last 3 years. There are some things that the Nokia does better than the iPhone ( I love the hardware button for the camera), and things I miss about the old phone (hardware mute button, combined email).

I've had this thing for a week and I'm already making apps. There are three of my creations on there right now! They are all really lame; they will make me laugh in a couple of months -- that's a start.

I like the iPhone, I like the apps, but it locked me into the apps that already exist (yes, I could a Mac and write apps in Objective-C; I have a job and other commitments). As a developer in the Microsoft space, the Windows Phone puts me into the "What can I get this thing to do" space. With the iPhone, it's "Where can I find an app to".

I am interested in building apps it get a feel of how they work than make my first million selling Winphone apps. I want to get a better feel for the mobile metaphors. If Winphone crashes and burns, I can transfer these skills to Droid, iOS or whatever comes next. Developing for Winphone allows me to concentrate on the smaller device and not deal with Java or Objective-C until after I get the new platform.

I think it is important to learn the new stuff. Whether its iPad, Windows 8, some Android based Tablet or something I don't yet know about, the tech world is changing and I need to understand it.

Saturday, March 24, 2012

Doing Notes for my Boise Code Camp Presentation

I presented this afternoon at the Boise Code Camp. All in all, it went OK. However, I had a really rough start. I am proud that I was able to brush it off and go on (I even got a couple of complements). It is important to continue on when I am bombing and debug my performance after the ordeal is over. After an Improv show (at least at the Blue Door Theatre), the performers do “Notes”.

I have observed that many code camp presentations start out scripted and eventually move into improv. It isn’t like a short games or a Harold where it can go in any direction; more like a Larry David script that has a story arc but no written dialog. This pivot usually occurs when the first question is asked. The presentation becomes a dialog that follows the outline of the presentation. I usually am good at engaging the audience in that dialog; I guess I am good at crowd work.

Rough Start

I had a lot of trouble getting started. I umed, I spoke too fast, spoke in a single note, I left stuff off (I didn’t even properly introduce myself), it took too long to get into a groove. My first 5 minutes were really weak. I didn’t rehearse that part of the presentation enough. I didn’t speak the words out loud enough. The first 5 minutes sets the pace for the rest of the show.

Improvements in Other Areas

I felt good about the rest of the talk. Last time out I had trouble talking and coding at the same time; this time I felt comfortable coding on stage. The steps of my demos were well defined and rehearsed; I even had completed demos that I could use if I had a demo failure. I also felt comfortable changing the demos based on questions or comments of my audience. I need to keep the improvements I made and fix what didn’t work

Perspective

It is not my goal to oust Scott Hanselman as the Jon Skeet of technical speaking. I am using technical speaking (and Improv for that matter) to help me get outside of myself and the traditional dweeb mindset. I am getting out of my shell and doing thing I wouldn’t have done and going places I wouldn’t have gone 3 years ago.