9.5.5. Numeric Conversions

Strings often contain characters that represent numbers. For example, we represent the numeric value 15 as a string with two characters, the character '1' followed by the character '5'. In general, the character representation of a number differs from its numeric value. The numeric value 15 stored in a 16-bit short has the bit pattern 0000000000001111, whereas the character string "15" represented as two Latin-1 chars has the bit pattern 0011000100110101. The first byte represents the character '1' which has the octal value 061, and the second byte represents '5', which in Latin-1 is octal 065.

The new standard introduced several functions that convert between numeric data and library strings:

int i = 42;
string s = to_string(i);  // converts the int i to its character representation
double d = stod(s);       // converts the string s to floating-point

Table 9.16. Conversions between strings and Numbers

Here we call to_string to convert 42 to its corresponding string representation and then call stod to convert that string to floating-point.

The first non-whitespace character in the string we convert to numeric value must be a character that can appear in a number:

string s2 = "pi = 3.14";
// convert the first substring in s that starts with a digit,  d = 3.14
d = stod(s2.substr(s2.find_first_of("+-.0123456789")));

In this call to stod, we call find_first_of (§ 9.5.3, p. 364) to get the position of the first character in s that could be part of a number. We pass the substring of s starting at that position to stod. The stod function reads the string it is given until it finds a character that cannot be part of a number. It then converts the character representation of the number it found into the corresponding double-precision floating-point value.

The first non-whitespace character in the string must be a sign (+ or -) or a digit. The string can begin with 0x or 0X to indicate hexadecimal. For the functions that convert to floating-point the string may also start with a decimal point (.) and may contain an e or E to designate the exponent. For the functions that convert to integral type, depending on the base, the string can contain alphabetic characters corresponding to numbers beyond the digit 9.