Question

I have noticed that a really cool method to convert a string, say

str = '1234'

to a vector is to use this trick.

vec = str - '0'
 = [1 2 3 4]

My question is why does this method work?

Further, something like:

vec1 = str -'1'
  = [0 1 2 3]

but

vec2 = str - '10'
  Error using  - 
  Matrix dimensions must agree.

What is taking place here?

Answer 1

When you use arithmetic operators with strings, Matlab casts the strings as doubles, which converts a string to ascii values:

>> double('1')
ans =
    49

Thus, subtraction will work just fine, though addition will give weird results

>> '1'+'1'
ans =
    98

Converting an array of strings to double results in an array of doubles, therefore the “matrix dimensions must agree”:

>> double('10')
ans =
    49    48

Thus, while subtracting '0' is thus a cool shortcut, I suggest you use STR2DOUBLE instead to avoid confusion.