Asp Core 1.1, DataTable exists in both

Question

I am developing a dot net core 1.1 app in which I am trying to use Accord.Net. According to examples in this page (Naive Bayes) I need to convert data retrieved from DB to DataTable.

The thing is that while using DataTable I got this error:

The type 'DataTable' exists in both 'Shim, ...' and 'System.Data.Common, ...'

Even if I use this:

DataTable learningDataNotCodifiedAsDataTable = new DataTable();

or this:

System.Data.DataTable learningDataNotCodifiedAsDataTable = new System.Data.DataTable();

TG.

Answer 1

If you have System.Data assembly in Assemblies and don't want or can't delete it, then you can bypass it by using extern alias, but when I bypassed this error using it I got 'DataTable' does not contain a constructor that takes 0/1 arguments error, and if believe this discussion the reason is:

System.Data.DataTable is present in .Net core(1.0,1.1) as an empty class to complete the interfaces implementation. This issue is to track the work needed to bring in an API to provide DataTable like API in .Net Core.

And it changed only in .NET Core 2.0, see this SO post. I tried you code in .NET Core 2.0 project (in VS 2017 15.3) and only then it worked fine.

UPDATE: I meant this assemblies.

But as you say you have only NUGET packages, then you also can use aliases in you csproj file for Nuget packages like below(I used System.Data.Common you can replace it with your Shim package if needed) :

 <Target Name="DataAlias" BeforeTargets="FindReferenceAssembliesForReferences;ResolveReferences">
    <ItemGroup>
      <ReferencePath Condition="'%(FileName)' == 'System.Data.Common'">
        <Aliases>MyData</Aliases>
      </ReferencePath>
    </ItemGroup>
  </Target>

and then reference it in C# like this:

extern alias MyData; //1st line in .cs file
...
using MyData::System.Data;
...
DataTable datatable = new DataTable();

But still you won't be able to use because you will get the error about constructor I wrote above. Here you has 2 options how to solve this:

1. Switch to .NET Core 2.0
2. Try to use workaround solution from this post using DbDataReader if it suits you

Answer 2

While the DataTable is not available in .NET Core 1.1, it is now available in .NET Core 2.0. If you can upgrade your project to .NET Core 2.0, then you will be able to use it in your code.

However, if you cannot switch to .NET Core 2.0 right now, then please note that you are not required to use DataTables with any of the methods in Accord.NET framework. They are given or shown just because they can give some extra convenience, but they are not really required, as shown in the example below:

string[] columnNames = { "Outlook", "Temperature", "Humidity", "Wind", "PlayTennis" };

string[][] data =
{
    new string[] { "Sunny", "Hot", "High", "Weak", "No" },
    new string[] { "Sunny", "Hot", "High", "Strong", "No" },
    new string[] { "Overcast", "Hot", "High", "Weak", "Yes" },
    new string[] { "Rain", "Mild", "High", "Weak", "Yes" },
    new string[] { "Rain", "Cool", "Normal", "Weak", "Yes" },
    new string[] { "Rain", "Cool", "Normal", "Strong", "No" },
    new string[] { "Overcast", "Cool", "Normal", "Strong", "Yes" },
    new string[] { "Sunny", "Mild", "High", "Weak", "No" },
    new string[] { "Sunny", "Cool", "Normal", "Weak", "Yes" },
    new string[] {  "Rain", "Mild", "Normal", "Weak", "Yes" },
    new string[] {  "Sunny", "Mild", "Normal", "Strong", "Yes" },
    new string[] {  "Overcast", "Mild", "High", "Strong", "Yes" },
    new string[] {  "Overcast", "Hot", "Normal", "Weak", "Yes" },
    new string[] {  "Rain", "Mild", "High", "Strong", "No" },
};

// Create a new codification codebook to
// convert strings into discrete symbols
Codification codebook = new Codification(columnNames, data);

// Extract input and output pairs to train
int[][] symbols = codebook.Transform(data);
int[][] inputs = symbols.Get(null, 0, -1); // Gets all rows, from 0 to the last (but not the last)
int[] outputs = symbols.GetColumn(-1);     // Gets only the last column

// Create a new Naive Bayes learning
var learner = new NaiveBayesLearning();

NaiveBayes nb = learner.Learn(inputs, outputs);

// Consider we would like to know whether one should play tennis at a
// sunny, cool, humid and windy day. Let us first encode this instance
int[] instance = codebook.Translate("Sunny", "Cool", "High", "Strong");

// Let us obtain the numeric output that represents the answer
int c = nb.Decide(instance); // answer will be 0

// Now let us convert the numeric output to an actual "Yes" or "No" answer
string result = codebook.Translate("PlayTennis", c); // answer will be "No"

// We can also extract the probabilities for each possible answer
double[] probs = nb.Probabilities(instance); // { 0.795, 0.205 }