[ad_1]
\n<\/p>\n
I was curious about the relative performance of these conversions so I did a few tests with the naive versions of a few methods. A few notes and caveats:<\/p>\n
std::to_string()<\/code> versions.<\/li>\n<\/ul>\nAlgorithms tested are:<\/p>\n
string toString1 (const int a)\n{\n char buffer[32];\n string result = itoa(a, buffer, 10);\n char lastDigit = result[result.length() - 1];\n result[result.length() - 1] = '.';\n result += lastDigit;\n\n return result;\n}\n\nvoid toString1a (string& Output, const int a)\n{\n char buffer[32];\n Output = itoa(a, buffer, 10);\n char lastDigit = Output[Output.length() - 1];\n Output[Output.length() - 1] = '.';\n Output += lastDigit;\n}\n\nstring toString2 (const int a) {\n float f = a * 0.1f;\n ostringstream ss;\n ss << f;\n\n return ss.str();\n} \n\nconst char* toString3 (const int a)\n{\n static char s_buffer[32]; \/\/Careful!\n\n itoa(a, s_buffer, 10);\n size_t len = strlen(s_buffer);\n char lastDigit = s_buffer[len-1];\n s_buffer[len-1] = '.';\n s_buffer[len] = lastDigit;\n s_buffer[len+1] = '\\0';\n\n return s_buffer;\n} \n\nconst char* toString4 (char* pBuffer, const int a)\n{\n itoa(a, pBuffer, 10);\n size_t len = strlen(pBuffer);\n char lastDigit = pBuffer[len-1];\n pBuffer[len-1] = '.';\n pBuffer[len] = lastDigit;\n pBuffer[len+1] = '\\0';\n\n return pBuffer;\n}\n\nvoid toString4a (char* pBuffer, const int a)\n{\n itoa(a, pBuffer, 10);\n size_t len = strlen(pBuffer);\n char lastDigit = pBuffer[len-1];\n pBuffer[len-1] = '.';\n pBuffer[len] = lastDigit;\n pBuffer[len+1] = '\\0';\n}\n\nconst char* toString5 (char* pBuffer, const int a)\n{\n snprintf(pBuffer, 16, \"%.1f\", a\/10.0);\n return pBuffer;\n}\n\nconst char* toString6 (char* pBuffer, const int a)\n{\n snprintf(pBuffer, 16, \"%d.%01d\", a \/ 10, a % 10);\n return pBuffer;\n}\n\nstring toString7 (const int a)\n{\n ostringstream stream;\n stream << (a \/ 10) << '.' << (a % 10);\n return stream.str();\n}\n\nstring toString8 (const int a)\n{\n char buffer[16];\n string result = _itoa(a, buffer, 10);\n result.insert(result.length()-1, 1, '.'); \n return result;\n}\n<\/code><\/pre>\nBasic benchmarks were run on the above methods for 100 million integers (starting from 0). If performance really matters for your case please profile\/benchmark for yourself for your exact use-case.<\/p>\n
\n- toString1 = 0.64 sec (itoa to string)<\/li>\n
- toString1a = 0.61 sec (inplace, no return)<\/li>\n
- toString2 = 33.4 sec (stringstream, ouch!)<\/li>\n
- toString3 = 0.52 sec (itoa, static buffer)<\/li>\n
- toString4 = 0.52 sec (itoa)<\/li>\n
- toString4a = 0.52 sec (itoa, inplace)<\/li>\n
- toString5 = 6.35 sec (snprintf, float)<\/li>\n
- toString6 = 2.29 sec (snprintf, two ints)<\/li>\n
- toString7 = 33.3 sec (stringstream, two ints)<\/li>\n
- toString8 = 0.70 sec (itoa and string::insert)<\/li>\n<\/ul>\n
Which one is “better” really depends on what you need. Personally I would use the simplest stream case (toString7) unless I was absolutely sure I needed the performance of the other methods.<\/p>\n<\/p><\/div>\n