Question

Lets say I have some known library to load data from an .xls file and returns a DataTable populated with data from the first sheet in any excel workbook. Also a Log function that prints out messages in the absence of a working debugger.

DataTable dtFoo = null;
DataTable dtBar = null;
DataTable dtChaz = null;

String[] files = new String[]{ "file1.xls", "file2.xls", "file2.xls" };
DataTable[] dts = new DataTable[] { dtFoo, dtBar, dtChaz };

for(int i = 0; i < 3; i++)
{
    dts[i] = SomeLibrary.LoadFromFile(files[i]); //Returns a new DataTable
    Log((dts[i] == null) + " " + dts[i].Rows.Count)
}

Log((dts[0] == null) + " " + (dtFoo == null));
Log((dts[1] == null) + " " + (dtBar == null));
Log((dts[2] == null) + " " + (dtChaz == null));

Log Output:

  False 40
  False 455
  False 34
  False True
  False True
  False True

Clearly I’m missing something important when working with reference type variables that I haven’t been able to figure out. Why are my DataTable variables still null after the loop completes?

Answer 1

This line:

DataTable[] dts = new DataTable[] { dtFoo, dtBar, dtChaz };

… doesn’t alias the variables with the array elements, as you seem to think it does. It just creates a new array, and sets the initial values to the values of dtFoo, dtBar and dtChaz respectively (all of which are null). The array and the variables are entirely separate afterwards.

If you want to set the variables later, you’ll need:

dtFoo = dts[0];
dtBar = dts[1];
dtChaz = dts[2];

Of course this doesn’t have to be with DataTables. Simple example:

string x = "initial";
string[] array = new string[] { x };
array[0] = "modified";

Console.WriteLine(x); // "initial"
x = array[0];
Console.WriteLine(x); // "modified"

Note that this is when you change the value of the array element. If instead you modify the object that the value refers to, that’s a different matter:

StringBuilder builder = new StringBuilder("123");
StringBuilder[] array = new StringBuilder[] { builder };
builder[0].Append("456");
Console.WriteLine(builder); // 123456

See my article on reference types and value types for more details. Basically you need to differentiate between references and the objects they refer to.