Inputname statement - useful example

Hi everyone,

could you help me to understand what is the useful of the inputname statement? I run the following example:

function c = somma(a, b)
  disp(inputname(1))
  c = a + b;

in workspace:

x = 1;
y = 2;
somma(x, y)            %x is the result

From my example is not clear the useful of inputname. If you have a best example, let me know please.

Thank you.

[Stephen23]

In general it is not useful: the names that variables have inside and outside of functions should be totally irrelevant from outside and inside respectively.

Answer 1

[Walter Roberson]

inputname() is useful for giving error messages that reference variable names in the form the user knowns them.

[Gennaro Arguzzi]

Hi @Walter Roberson, could you show me a simple example please?

[Walter Roberson]

function tryme(var1)
  if ~exist('var1', 'variable') || isempty(var1) || ndims(var1) ~= 3
    varname = inputname(1);
    if isempty(varname); varname = 'first input';
      error('%s must be a 3D array', varname);
    end
  end

Now tryme with an expression will talk about "first input", but if it is passed a variable name, the variable name will be referred to.

[Gennaro Arguzzi]

Walter, unfortunately matlab shows the following errors:

tryme(1)

Error using exist The optional second input to exist must be 'var', 'builtin', 'class', 'dir' or 'file'.

Error in tryme (line 2) if ~exist('var1', 'variable') isempty(var1) ndims(var1) ~= 3

[Walter Roberson]

function tryme(var1)
  if ~exist('var1', 'var') || isempty(var1) || ndims(var1) ~= 3
    varname = inputname(1);
    if isempty(varname); varname = 'first input'; end
    error('%s must be a 3D array', varname);
  end

Answer 2

[Jan]

The command inputname is not useful. The name of a variable should not matter. Some inputs do not even have a name, e.g. in: somma(1, 2) or somma(x+1, y+2).

There are some rare situations, where the name is useful, e.g. when you want to display a variable.

I'm programming for many years now and was never faced with the need of using inputname. If the name of a value matters, I create a struct, which carries the name in an extra field.

Answer 3

[dpb]

I don't know as I have a directly useful real example on hand, and in general functions shouldn't rely on needing outside knowledge for their use, but one could imagine places in which what the function should do depends on what the variable passed is.

A totally generic routine for reading some data could by this mechanism create a labeled output of the inputs if it were called with different arguments at different points in the code. One could envision this particularly in cases using the dreaded eval in the calling routine so that the calling routine actually is also variable.

I would NOT recommend this as a common practice; it's there for those rare situations that might call for it.

[Stephen23]

"One could envision this particularly in cases using the dreaded eval in the calling routine"

The inputname documentation explicitly warns against this: "Avoid using inputname in the critical path of code or to obtain variable names to be used with commands such as eval, evalin, and assignin. This use of inputname can lead to code that is difficult to maintain."

[dpb]

Oh, yeah, did say it is "the dreaded eval" and I meant that explicitly, NOT tongue-in-cheek. I was just trying to come up with a reason why one could need to retrieve a name from the caller in the callee, not recommending same.