问题
For a task, I'm doing some simple data sampling to determine which samples contains audio counting the total number of energy. I've been looking into the std::count_if
function, and, although this suits my needs to a certain extent, for example:
int foo = std::count_if(
std::begin(samples),
std::end(samples),
containsSound);
This counts the total number of samples that contain sound, but does not give an indication to the samples that contain sound. I came up with this solution:
std::vector<std::string> sound_files = myLib::count_sample_if(
std::begin(samples),
std::end(samples),
samples.DirName,
containsSOund);
This would then store and push the samples.DirName
to a vector
which I can then use to only store the sample set of my choice.
My question is whether or not this would be easy to implement?
回答1:
If you just need readability / speed of development and you don't care about performance then you can easily use std::copy_if
and std::transform
to obtain what you need:
std::vector<Song> songs;
std::vector<Song> filtered;
std::vector<std::string> transformed;
std::copy_if(songs.begin(), songs.end(), filtered.begin(), [](const Song &song) { return whatever you need; });
std::transform(filtered.begin(), filtered.end(), transformed.begin(), [](const Song &song) { return song.sample; });
Or you could use std::for_each
:
std::vector<Song> songs;
std::vector<std::string> transformed;
std::for_each(songs.begin(), songs.end(), [&](const Song &song) { if (song.containsSample()) transformed.push_back(song.sample); });
Then the amount of samples that contains sound is just transformed.size()
.
回答2:
If I have understood correctly the simplest way is to use standard algorithm std::copy_if. There is no need to count elements because you can simply get it by using another standard function std::distance. For example let assume that you have an array of integers and want to count positive values and at the same time to copy them in a vector. The code could look the following way
int a[] = { 1, -3, -5, 9, 2, -4, -1, -7, 5, 8 };
std::vector<int> v;
std::copy_if( std::begin( a ), std::end( a ),
std::back_inserter( v ),
std::bind2nd( std::greater<int>(), 0 ) );
std::cout << "The number of positive elements is " << std::distance( v.begin(), v.end() ) << std::endl;
回答3:
It would be very easy to implement - additionally, there is no need to implement it. You can write a functor (aka function object) or lambda expression to do the comparison and hold the vector for you, which would allow you to continue to use std::count_if
(or one of the other standard algorithms).
std::vector<std::string> mySongs;
std::string myDirectory = "C:/";
std::copy_if(std::begin(samples), std::end(samples), std::back_inserter(mySongs), [](const std::string& s)
{
// return true or false based on some criteria here
});
std::transform(std::begin(mySongs), std::end(mySongs), std::begin(mySongs), [](const std::string& s)
{
return myDirectory + s;
});
回答4:
here is the implementation of count_if, all you need to do is to create a vector, push back the result if it is acceptable, and return the vector after you finish looping through the range.
来源:https://stackoverflow.com/questions/20568278/c-developing-own-version-of-stdcount-if