我正在尝试编写一个计算密集型程序。我需要char *作为multi_index_container的composite_key_compare的comparsion字段。但是,它似乎不起作用。代码如下:
struct MyStruct
{
char* firstName;
char* secondName;
int age;
};
struct equal_char
{ // functor for operator<=
inline bool operator()(const char* left, const char* right) const
{ // apply operator<= to operands
bool result=(strcmp(left,right)==0);
return result;
}
};
typedef composite_key
<MyStruct*,
BOOST_MULTI_INDEX_MEMBER(MyStruct, char*, firstName),
BOOST_MULTI_INDEX_MEMBER(MyStruct, char*, secondName)
> comp_key;
typedef multi_index_container
<
MyStruct*,
indexed_by
<
ordered_unique
<
comp_key,
composite_key_compare
<equal_char, equal_char>
>
>
> MyContainer;
boost::ptr_vector<MyStruct> vec;
MyStruct* struct1=new MyStruct();
struct1->firstName="Michael";
struct1->secondName="Mike";
struct1->age=20;
vec.push_back(struct1);
MyContainer myContainer;
myContainer.insert(struct1);
char* first="Michael";
char* second="Mike";
auto it=myContainer.find(boost::make_tuple(first, second));
if(it!=myContainer.end())
cout << (*it)->age << endl;
我确实追踪到了equal_char,发现它确实在“Michael”与“Michael”的第一次比较中返回true,但我也发现equal_char没有被称为“Mike”的第二次比较“麦克风”。谁可以帮我这个?我该如何编写composite_key_compare?
答案 0 :(得分:0)
我跟踪了程序,发现对于“ordered unique”的比较操作,boost需要类似std :: less()的东西。
所以,我在下面的代码中添加了它,并且它有效。
struct CompareLess
{ // functor for operator<=
#pragma region For 1 variable
static inline int compare(const char* left, const char* right)
{
return strcmp(left, right);
}
inline bool operator()(const char* left, const char* right) const
{ // apply operator<= to operands
return compare(left, right)<0;
}
static inline int compare(const boost::tuple<char*>& x, const char*y)
{
return compare(x.get<0>(),y);
}
inline bool operator()(const boost::tuple<char*>& x, const char*y) const
{
return compare(x,y)<0;
}
static inline int compare(const boost::multi_index::composite_key_result<comp_key>& k, const boost::tuple<char*>& y)
{
return -compare(y,(const char*)(k.value->firstName));
}
inline bool operator()(const boost::multi_index::composite_key_result<comp_key>& k, const boost::tuple<char*>& y) const
{
return compare(k,y)<0;
}
static inline int compare(const boost::tuple<char*>& y, const boost::multi_index::composite_key_result<comp_key>& k)
{
return compare(y,(const char*)(k.value->firstName));
}
inline bool operator()(const boost::tuple<char*>& y, const boost::multi_index::composite_key_result<comp_key>& k) const
{
return compare(y,k) <0;
}
#pragma endregion For 1 variable
#pragma region For 2 variables
static inline int compare(const boost::multi_index::composite_key_result<comp_key>& k, const boost::tuple<char*, char*>& y)
{
int val=strcmp(k.value->firstName, y.get<0>());
if(val!=0)
return val;
else
return compare(y.get<1>(),(const char*)(k.value->secondName));
}
inline bool operator()(const boost::multi_index::composite_key_result<comp_key>& k, const boost::tuple<char*, char*>& y) const
{
return compare(k,y) <0;
}
static inline int compare(const boost::tuple<char*, char*>& y, const boost::multi_index::composite_key_result<comp_key>& k)
{
return -compare(k,y);
}
inline bool operator()(const boost::tuple<char*, char*>& y, const boost::multi_index::composite_key_result<comp_key>& k) const
{
return compare(y,k)<0;
}
#pragma endregion For 2 variables
#pragma region For all variables
inline bool operator()(const boost::multi_index::composite_key_result<comp_key>& k1,
const boost::multi_index::composite_key_result<comp_key>& k2) const
{
return CompareLess()((const char*)(k1.value->firstName), (const char*)(k2.value->firstName));
}
#pragma endregion For all variables
};
typedef multi_index_container
<
MyStruct*,
indexed_by
<
ordered_unique
<
comp_key,
CompareLess
>
>
> MyContainer;
但是我得到了更多的问题,这是一个非常冗长的代码,并且假设如果我有更多的字段,它将会更长。有没有聪明的方法来做到这一点?