Item 5: Prefer range member functions to their single-element counterparts
1. It's generally less work to write the code using the range member functions. Or you'd have to write an explicit loop, and the loop also happens to impose an efficiency penalty.
2. Range member functions tend to lead to code that is clearer and more straightforward.
3.Almost all uses of copy where the destination range is specified using an insert iterator should be replaced with calls to range member functions, because the copy alogrithm is actually using a single-element member functions.
4. When dealing with the standard sequence containers, application of single-element member functions makes more demands on memory allocators, copies objects more frequently, and/or performs redundant operations compared to range member functions that achieve the same end.
The following analysis applies for vector, string, and deque
a.The first tax consists of unnecessary function calls.
b.the second tax, which is the cost of inefficiently moving the existing elements in v to their final post-insertion positions.In contrast, the Standard requires that range insert functions move existing container elements directly into their final positions.
c.The final performance tax levied on those so foolish as to use repeated single-element insertions instead of a single range insertion has to do with memory allocation, though it has a nasty copying side to it, too
As to list, first item marked as 'a' still stands, and a new problem for single-element member functions is repeated superfluous assignments to the next and prev pointers of some nodes in the list.
For the associative containers, the efficiency case is harder to make, though the issue of extra function call overhead for repeated calls to single-element insert continues to apply. Furthermore, certain special kinds of range insertions may lead to optimization possibilities in associative containers, too, but as far as I can tell, such optimizations currently exist only in theory. By the time you read this, of course, theory may have become practice, so range insertions into associative containers may indeed be more efficient than repeated single-element insertions. Certainly they are never less efficient, so you have nothing to lose by preferring them.
2. Range member functions tend to lead to code that is clearer and more straightforward.
3.Almost all uses of copy where the destination range is specified using an insert iterator should be replaced with calls to range member functions, because the copy alogrithm is actually using a single-element member functions.
4. When dealing with the standard sequence containers, application of single-element member functions makes more demands on memory allocators, copies objects more frequently, and/or performs redundant operations compared to range member functions that achieve the same end.
The following analysis applies for vector, string, and deque
a.The first tax consists of unnecessary function calls.
b.the second tax, which is the cost of inefficiently moving the existing elements in v to their final post-insertion positions.In contrast, the Standard requires that range insert functions move existing container elements directly into their final positions.
c.The final performance tax levied on those so foolish as to use repeated single-element insertions instead of a single range insertion has to do with memory allocation, though it has a nasty copying side to it, too
As to list, first item marked as 'a' still stands, and a new problem for single-element member functions is repeated superfluous assignments to the next and prev pointers of some nodes in the list.
For the associative containers, the efficiency case is harder to make, though the issue of extra function call overhead for repeated calls to single-element insert continues to apply. Furthermore, certain special kinds of range insertions may lead to optimization possibilities in associative containers, too, but as far as I can tell, such optimizations currently exist only in theory. By the time you read this, of course, theory may have become practice, so range insertions into associative containers may indeed be more efficient than repeated single-element insertions. Certainly they are never less efficient, so you have nothing to lose by preferring them.
浙公网安备 33010602011771号