If you have a char that is in the range \'0\' to \'9\' how do you convert it to int values of 0 to 9
And then how do you convert it back?
Also given letters
There are two main ways to do this conversion: Lookup and Mathmatically
All ASCII values are denoted in decimal notion in this answer
Note that in ASCII: '0' is 48, 'A' is 65, and 'a' is 97
Lookup:
In the lookup version you have an array of char, and then place the mapped values in the array, and create an array of ints to convert back:
In order to both validate and get the corresponding value when mapping char to int:
0 will be a sentinal value to mean not mapped: out of range
all results will be one more than expected
unsigned char is used to make sure a signed negative char is handled correctly
While 'C' allows the notation { ['A'] = 1, ['B'] = 2,… }; , C++ does not, so generically the following code can be used to fill lookup tables:
void fill_lookups(unsigned char * from_table, int from_size, int * to_table)
{
for (int i = 0; i < from_size; ++i)
{
to_table[from_table[i]]=i+1; // add one to support 0 as "out of range"
}
}
unsigned char int_to_char[]={ '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' };
unsigned char int_to_lower[]={'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j',
'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't',
'u', 'v', 'w', 'x', 'y', 'z'};
unsigned char int_to_upper[]={'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J',
'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T',
'U', 'V', 'W', 'X', 'Y', 'Z'};
int char_to_int[UCHAR_MAX+2] = {}; // This will return 0 for non digits
int letter_to_offset[UCHAR_MAX+2] = {}; // This will return 0 for non alpha
fill_lookups(int_to_char, sizeof(int_to_char), char_to_int);
fill_lookups(int_to_lower, sizeof(int_to_lower), letter_to_offset);
fill_lookups(int_to_upper, sizeof(int_to_upper), letter_to_offset);
// Helper function to check in range and always reduce in range lookups by 1
int to_int(int * table, unsigned char c, bool * in_range)
{
int ret = table[c];
if (ret)
{
*in_range=(1==1); // for C/C++ true
--ret;
}
else
{
*in_range=(0==1); // for C/C++ false
}
return ret;
}
bool in_range; // always true in these cases
int a=to_int(char_to_int, '7', &in_range); // a is now 7
char b=int_to_char[7]; // b is now '7'
int c=to_int(letter_to_offset, 'C', &in_range); // c=2
int d=to_int(letter_to_offset, 'c', &in_range); // d=2
char e=int_to_upper[2]; // e='C'
char f=int_to_lower[2]; // f='c'
While this will work and if validation or other lookups are needed this might make sense, but...
In general a better way to do this is using mathmatic equations.
Mathmatically (alpha works for ASCII)
Assuming that the conversions have already been validated to be in the correct range: (C style cast used for use with C or C++)
Note that '0'-'9' are guarenteed to be consecutive in C and C++
For ASCII 'A-Z' and 'a-z' are not only consecutive but 'A' % 32 and 'a' % 32 are both 1
int a='7'-'0'; // a is now 7 in ASCII: 55-48=7
char b=(char)7+'0'; // b is now '7' in ASCII: 7 + 48
int c='C' % 32 - 1; // c is now 2 in ASCII : 67 % 32 = 3 - 1 = 2
-or- where we know it is uppercase
int c='C'-'A'; // c is now 2 in ASCII : 67 - 65 = 2
int d='c' % 32 - 1; // d is now 2 in ASCII : 99 % 32 = 3 - 1 = 2
-or- where we know it is lowercase
int d='c'-'a'; // d is now 2 in ASCII : 99 - 97 = 2
char e=(char)2 + 'A'; // e is 'C' in ASCII : 65 + 2 = 67
char f=(char)2 + 'a'; // f is 'c' in ASCII : 97 + 2 = 99