My use case:
map cars;
bool exists(const string& name) {
// somehow I should find whether my MAP has a car
// with the name provid
std::map::find(const key_type& x );
It returns map::end
if the item doesn't exist.
bool exists(const std::map<std::string, Car>& cars, const std::string& name) {
return cars.end() != cars.find(name);
}
What about:
template <typename KeyType, typename Collection>
bool exists_in(Collection const& haystack, KeyType const& needle) {
return std::find(haystack.begin(), haystack.end(), needle) != haystack.end();
}
template <typename K, typename V>
bool exists_in(std::map<K,V> const& haystack, K const& needle) {
return haystack.find(needle) != haystack.end();
}
This makes exists_in
work with any standard container via std::find
and use a special version for std::map
since it offers a more efficient searching alternative. You could add additional specializations as necessary (e.g., for std::set
and others).
bool exists(const string& name)
{
return cars.find(name) != cars.end();
}
Apart from the answers with iterator-Value from find() and comparison to .end(), there is another way: map::count.
You can call map::count(key) with a specific key; it will return how many entries exist for the given key. For maps with unique keys, the result will be either 0 or 1. Since multimap exists as well with the same interface, better compare with != 0 for existence to be on the safe side.
for your example, that's
return (cars.count(name)>0);
The advantages I see are 1. shorter code, 2. benefit from whatever optimisations the library may apply internally, using its representation details.
#define itertype(v) typeof((v).begin())
itertype(cars) it = cars.find(name);
return it != cars.end();