不管從哪個角度看，切割字串都不是件簡單的工作
即便是簡單的字串格式
The simplest Example
先來個C語言處理方式
1
2
3
4
5
6
7
8
9
| #include <stdio.h>
int main()
{
char str[] = "Tom 42";
int value;
char name[30];
sscanf(str, "%s %d", name, &value);
printf("%s %d\n", name, value);
}
|
C++的版本
1
2
3
4
5
6
7
8
9
10
11
12
| #include <sstream>
#include <iostream>
int main()
{
std::string str = "Tom 42";
int value;
std::string name;
std::stringstream ss;
ss << str;
ss >> name >> value;
std::cout << name << " " << value << "\n";
}
|
看起來相差無幾，也沒需要動用到其他武器的地方
不過如果情況更複雜該怎麼辦
More Complex issue
例如輸入的字串是Tom: 42
如果什麼都不改的話，name會得到Tom:，不是我們想要的
如果改C語言的版本Parse string的字串
1
| sscanf(str, "%s: %d", name, &value);
|
結果也不是我們想要的
看看C++的版本，不改的話結果一樣
如果改成這樣
1
2
3
| int value;
std::string name, unused;
ss >> name >> unused >> value;
|
很顯然地也不對
Strtok
如果不動用重型武器，這是我唯一想得出來的方法
1
2
3
4
5
6
7
8
9
10
11
12
13
| #include <stdio.h>
#include <string.h>
#include <stdlib.h>
int main()
{
char str[] = "Tom: 42", *p = strtok(str, ":");
int value;
char name[30];
strcpy(name, p);
p = strtok(NULL, ":");
value = atoi(p);
printf("%s %d\n", name, value);
}
|
雖然解決了問題，不過程式碼支離破碎，維護起來也是麻煩透頂
萬一字串變成 Tom: 42, 123該怎麼辦
Boost Spirit
這世上不缺聰明人，想到了優雅到爆炸的作法
在C++內嵌DSL Parser解決問題，而使用的就是Boost Spirit
不過要說缺點的話，出錯Debug很麻煩
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
| #include <string>
#include <boost/spirit/include/qi.hpp>
namespace qi = boost::spirit::qi;
int main()
{
std::string str = "Tom: 42";
typedef std::string::const_iterator It;
It iter = str.begin(), end = str.end();
int value;
std::string name;
bool r = qi::phrase_parse(iter, end,
(*(qi::char_ - ':') >> ":" >> qi::int_), qi::ascii::blank,
name, value);
if (r && iter == end) {
std::cout << "Parsing succeeded\n";
std::cout << name << " " << value << "\n";
}
else {
std::cout << "Parsing failed\n";
std::cout << "stopped at: \"" << std::string(iter, end) << "\"\n";
}
}
|
就如同上面說的如果要修改字串成Tom: 42, 123
我們可以將Parser寫成
1
2
3
| bool r = qi::phrase_parse(iter, end,
(*(qi::char_ - ':') >> ":" >> qi::int_ >> "," >> qi::int_), qi::ascii::blank,
name, value, value1);
|
高明的不得了
Rule & Grammer
當要Parse的字串越來越複雜，就可以自己定義Grammer和Rule了
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
| #include <string>
#include <boost/tuple/tuple.hpp>
#include <boost/spirit/include/qi.hpp>
#include <boost/fusion/include/boost_tuple.hpp>
using MyTuple = boost::tuple<std::string, int, int>;
namespace qi = boost::spirit::qi;
template <typename Iterator, typename Skipper = qi::ascii::blank_type>
struct StringGrammar : qi::grammar <Iterator, MyTuple(), Skipper> {
StringGrammar() : StringGrammar::base_type(parser, "StringParser Grammar") {
name = *(qi::char_ - ':');
value = qi::int_;
parser = name >> ":" >> value >> "," >> value;
}
private:
qi::rule<Iterator, MyTuple(), Skipper> parser;
qi::rule<Iterator, std::string(), Skipper> name;
qi::rule<Iterator, int(), Skipper> value;
};
int main()
{
std::string str = "Tom: 42, 123";
typedef std::string::const_iterator It;
StringGrammar<It> grammer;
It iter = str.begin(), end = str.end();
MyTuple v;
bool r = qi::phrase_parse(iter, end,
grammer, qi::ascii::blank,
v);
if (r && iter == end) {
std::cout << "Parsing succeeded\n";
std::cout << boost::get<0>(v) << " " << boost::get<1>(v) << " " << boost::get<2>(v) << "\n";
}
else {
std::cout << "Parsing failed\n";
std::cout << "stopped at: \"" << std::string(iter, end) << "\"\n";
}
}
|
經過測試，如果省略後面Skipper的模板定義，整個Grammer不會正常運作
Reference
– Home of The Boost.Spirit Library
– Parsing with Spirit Qi
– spirit.Qi in boost