Reputation: 3845
For instance. I have some structure:
s_Some{
std::string lable;
s_some_junk some_junk;
};
And a vector:
std::vector<s_Some> mSome;
And then I fill this vector with a lot of s_Somes.
I need to find an iterator for a single s_Some in this vector, which has a specific lable. So far I just iterate through all of this junk and match every lable with the one wanted. This looks a little bit stupid to me. Is there a better way to do so?
Upvotes: 3
Views: 13952
Reputation: 73443
You can find_if algorithm to do this. Define a predicate some thing like this:
struct isEqual
{
isEqual(const std::string& s): m_s(s)
{}
bool operator()(S_Some& l)
{
return l.lable == m_s;
}
std::string m_s;
};
And while searching you can use
std::vector<S_Some>::iterator iter = std::find_if(mSome.begin(),
mSome.end(),
isEqual(std::string("AAAA"));
Upvotes: 1
Reputation: 8926
Use a
std::multimap< string, s_Some > mSome;
and
mSome.insert( std::make_pair( aSome.lable, aSome ) );
Find the first instance of the value of lable you want by
mSome.find( lable_you_want );
The next instance will be found by incrementing the iterator.
cf. http://www.cppreference.com/wiki/stl/multimap/start
That is, unless you're required to use a std::vector.
Upvotes: 3
Reputation: 36828
If you are to search only a few times or if your vector is likely to have different content every time you search, there's unfortunately no alternative; you will have to iterate through the whole vector.
If however your vector is not going to change once created and you have to run a large number of searches, do this:
This will be far quicker.
Upvotes: 5
Reputation: 18631
Option 1) If you are compelled to use the std::vector, but once the vector is filled it stays unchanged, then you could sort the vector and use the binary search. The only cost would be the sorting then and there will be no additional overhead. Searching time is logarithmic O(logN).
Option 2) If you have the freedom and can choose different data structure, then consider using the map (also logarithmic) or unordered_map ( expected O(1), worst O(n) ).
I have just noticed that you said you wanted to match every label with the one being looked for. So I conclude you can have duplicate labels. Then for point 2 use corresponding multi_map containers, while for point 1 things get a bit messier.
Upvotes: 10