Why is isstr not recommended?

When using isstr in my code, I get a warning that it is not recommended and I should use ischar instead. But ischar doesn't provide the same functionality as isstr. For a string, ischar returns 0 and isstr returns 1, while for a character array, ischar returns 1 and isstr returns 1. There doesn't seem to be any replacement for isstr that provides the same functionality. Why is this warning displayed?

Answer 1

[Star Strider]

The string (link) datatype was introduced in R2016b, along with a number of functions to work with it, of which isstring (link) is one. I suspect the recommendations for isstr changed to be compatible with the new datatype, and avoid any ambiguity.

[Josh G.]

I wasn't aware of isstring, but that looks like a better replacement. I wonder why Mathworks recommends ischar though instead of isstring.

[Star Strider]

I don’t believe MathWorks is actually ‘recommending’ one over the other, just doing its best to avoid ambiguity. It very much depends on what you’re testing.

Example —

s = "string!";
c = 'character!';

ischar_test = [ischar(s) ischar(c)]
isstring_test = [isstring(s) isstring(c)]
isstr_test = [isstr(s) isstr(c)]

produces:

ischar_test =
  1×2 logical array
   0   1

isstring_test =
  1×2 logical array
   1   0

isstr_test =
  1×2 logical array
   0   1

This also illustrates the problem with isstr.