c++ Windows Form app: How to grant public access to DataTable object?

[Lost User]

You should declare the object in your Form class and initialise it in the constructor or FormLoad event. It is then available to every other event handler or method within that class, whether public or private. This is basic C++ and nothing specific to CLI or Form apps.

[User 11196121]

Cheers Richard. Please very quick question: do you know why the below code compiled but tirggered an error msg box

Unable to cast object of type System.Data.DataColumn to type System.IConvertible

lblvol->Text=System::Convert::ToString(sqrt(System::Convert::ToDouble(dbdataset->Columns["DataReturn"])));

I was just aiming at computing the standard deviation of my return series held in ["DataReturn""]. can't see anything wrong as sqrt () is a built-in function from

#include "math.h"

Cheers

[User 11196121]

alright.cheers.

[Lost User]

My first suggestion would be: "don't write compound statements like that". As you can see they are almost impossible to figure out when they go wrong. Use interim values so you can check each stage of the code, like:

string dataReturn = dbdataset->Columns["DataReturn"]; // assuming this item is a string value
double doubleValue = System::Convert::ToDouble(dataReturn); // convert to double, but check for errors
double squareRoot = sqrt(doubleValue);
lblvol->Text = Convert::ToString(squareRoot);

Easier and clearer, and simple to step through with the debugger when errors occur.

[User 11196121]

Richard, Thanks for the advice. I will try it and let you know. Any idea why string type does not work from my side as opposed to your post? I am currently going it to see if people have encountered a similar issue. Best,

[Lost User]

Member 11230372 wrote:

Any idea why string type does not work from my side

Not without seeing the context I'm afraid. You can only use a string in the place where a string is valid.

[User 11196121]

might be a missing #include... as something is not logic here. as a matter of fact while typing any type double, float, char they are highlighted in blue therefore you could define any variable anywhere in the code based on those type. however while typing string, the latter does not seem to be recognized at all (not even highlighted). string is considered as plain text and thus unrecognized bizzare... Cheers

[Lost User]

double, float, char are basic types which are part of the language, while string does not exist, but String (with capital S) is a class. If I spelt it with a lower case s before, then I apologise, as that is C# rather than C++.

[User 11196121]

Thanks Richard for the feedback, and no worries.

Quote:http://www.cplusplus.com/reference/string/string/?kw=string[^]

The string class is an instantiation of the basic_string class template that uses char (i.e., bytes) as its character type, with its default char_traits and allocator types (see basic_string for more info on the template).

I will therefore work on your previous post and revert back to you tomorrow. Cheers

[Lost User]

Yes, but you are working in C++/CLI (which is why one responder told you to use the correct forum), where the correct class is http://msdn.microsoft.com/en-us/library/system.string(v=vs.110).aspx?cs-save-lang=1&cs-lang=cpp#code-snippet-1[^].

[User 11196121]

Hi Richard Thanks. I've made a simple console app and can confirm that the String class does work and you could create string variables such as string codeproject without problem. However,based on facts it looks like a different story when dealing with c++/cli. Back to the main standard deviation calculation, i think it could not work as a columns can't directly be assigned to a type and treated like a variable (see below). it could have worked in vba by selecting the column range and perform the built-in functions: c++ sqrt() or math::sqrt(). Hence the solution of last resort: iterate through each row and perform the calculation in several steps (get the mean from "DataRet" column first, then carry out ... ) unless a you have a better idea. Cheers

Error C2440: 'initializing' : cannot convert from 'System::Data::DataColumn ^' to 'double'

DataTable^dbdataset=gcnew DataTable();
DataColumn^rt_col=gcnew DataColumn();
rt_col->ColumnName = "DataRet";
rt_col->DataType=System::Type::GetType("System.Double");
dbdataset->Columns->Add(rt_col);

                 sda->Fill(dbdataset);

                 double varet = dbdataset->Columns\["DataRet"\]; //already double type, no need System::Convert::ToDouble
                 double vola = sqrt(varet);
                 lblvol->Text = Convert::ToString(vola);

[Lost User]

Member 11230372 wrote:

it looks like a different story when dealing with c++/cli

Many things are different if you use C++/CLI, it is not something I have worked with, as C# is so much more intuitive, even though my background is mainly C++.

Member 11230372 wrote:

Error C2440: 'initializing' : cannot convert from 'System::Data::DataColumn ^' to 'double'

Of course you cannot convert a DataColumn to a double, they are totally different things; remember a column is a collection of cells not a single item. You can hold a double value in a DataColumn's cell, but you need to access that with both the row and column indices.

[User 11196121]

Thanks Richard. i've therefore coded the calculation of standard deviation (see below). However results differ from Excel from the 3rd decimal point.Do you have any idea, as i have done it meticulously? I remembered someone working on sql server told me that depending on datatype used, computing 1/2 lead to a value close to 0.5 but still not 0.5. Do you think it might be an explanation, here? Cheers

int obs;
double mu, Ssigma, Svari, sum=0,sumsqr=0;

                 for (int i=1;i-1<dbdataset->Rows->Count-1;i++){
                     dbdataset->Rows\[i-1\]\["DataRet"\]=(System::Convert::ToDouble(dbdataset->Rows\[i-1\]\["Population"\])/System::Convert::ToDouble(dbdataset->Rows\[i\]\["Population"\])-1);
                     sum = sum + System::Convert::ToDouble(dbdataset->Rows\[i-1\]\["DataRet"\]);
                     obs = (dbdataset->Rows->Count-1);
                     mu = sum /obs;
                     sumsqr=sumsqr + pow((System::Convert::ToDouble(dbdataset->Rows\[i-1\]\["DataRet"\])-mu),2);
                     Svari=sumsqr/(obs-1);
                     Ssigma=sqrt(Svari);
                     lblvol->Text = Convert::ToString(Ssigma);

[Lost User]

Sorry, I only ever spent a year doing advanced maths, and that was more than 50 years ago. :sigh:

[User 11196121]

No worries Richard, and please note that I do thank you for your overall support. Cheers