I have an array of structures, sorted on a member of a structure, something like:
struct foo
{
    int bar;
    double baz;
};
// An array of foo, sorted on .bar
foo foos[] = { ........ };
// foos[0] = {0, 0.245}
// foos[1] = {1, -943.2}
// foos[2] = {2, 304.222}
// etc...
I want to find the element with a specific .bar value. It might or might not be in the array, and I'd like to do it in O(log(n)) time, since the array is sorted.
std::lower_bound is what I'd normally go for, but I need to specify a comparison function. However, the type of the members of the array (struct foo) and the searched value (int) are not the same, thus, my comparator is:
bool comp(foo a, int b)
{
    // ...
}
// --- or ---
bool comp(int a, foo b)
{
    // ...
}
It looks like the first will work with gcc, but I was wondering if the order of the arguments to the comparison function was specified by the standard, or if I'm relying on compiler behavior.
I'd like to avoid constructing a foo to pass to std::lower_bound here, as a full foo isn't required, and could be costly. My other option would be to wrap the foo * in a custom iterator that only exposed the .bar member.
From the standard, 25.3.3.1/3, on std::lower_bound():
Returns: The furthermost iterator
iin the range[first, last]such that for any iteratorjin the range[first, i)the following corresponding conditions hold:*j < valueorcomp(*j, value) != false.
From that, you may use
bool comp(foo a, int b)
or you may compare two foo instances and then access bar in both of them.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With