C++ Why diff EOF checks recommended for text vs numeric?

My textbook recommends using the member accessor method iStreamVar.eof() when dealing with textual data and while (iStreamVar) when dealing with numeric data.

Can someone please explain why it would matter?

Quote from book:

Using the function eof to determine the end-of-file status works best if the input is text. The earlier method of determining the end-of-file status works best if the input consists of numeric data.

That is the only thing mentioned on the topic. After this, it just explains how the process works.

Answer

Which method you use for determining the end of data depends on how you use it. My guess is, both methods which your textbook mentions are used wrong, so they fail in different situations. That’s why it recommends using different methods in different situations.

The correct method is not trivial, and it depends on how important error resilience is for you.

If you want to read a space-delimited stream with numbers in it, and you are sure the file contains no errors, the code is simplest:

int value;
while (iStreamVar >> value)
{
   ...
}

Note that it’s not any of the two original options.

If your file contains space-delimited textual data, and you are sure there are no errors, use the same code (but declare the temporary variable as string instead of int).

If you want to detect and recover from errors, use more elaborate code. But I cannot recommend you any specific code structure – it depends on what exactly you want to do in case of errors. Also:

  • Are text records delimited by space or newline?
  • What if the input text-file contains an empty line?
  • Numbers – floating-point or not?
  • Numbers – if there is a stray character like a among number data, what to do?

So there is no single correct recipe for doing proper input with error resilience.