Thread: Which method is more efficient?

Always use size method. The vector already increments internally an integer for you and the size method returns it.

There is no rule that says a vector keeps its size in a data member that is returned by size().

To answer the original question, do whatever produces the correct results for your program in the simplest manner possible. Worrying about efficiency at this stage is called "premature optimisation" for a reason. So don't worry about it.

Unless your program has exhibited a particular performance problem and you have used a profiler to identify a particular section of code as the cause, you have no reason to prefer one method over the other. Odds are, whichever option you pick to "optimise" performance, unless you have used a profiler, will be wrong.

Originally Posted by Elkvis

Statistically speaking, the OP has an equal chance of being right or wrong, since it's a simple "is it more efficient?" sort of question.

You're assuming all that programmers do is flip a coin. In practice, programmers who are trying to optimise use invalid criteria for deciding what to optimise, so tend to skew the odds against themselves.

Really if .size() was not efficient enough and would cause you to need to maintain the size separately, then it would defeat the purpoe of having a .size() member.
The same argument doesn't quite work for std::list though, unfortunately. They left that one open to being O(n) due to other advantages that has.

Originally Posted by Elkvis

but a programmer such as the OP, who doesn't have a preconceived notion of the right way, is equally likely to choose the right or the wrong option. you'll note that I didn't say "all programmers." I said "the OP."

And I disagree.

The OP has asked which technique is more efficient. That is a particular mindset. There is a whole bunch of issues around trying to maximise efficiency, such as it being hard to measure, variation (for a given code construct) between compiler and library implementations, and the "obvious" rules of thumb often being wrong (or, at best, implementation specific) in practice.

By asking the wrong question, and assuming there is only one answer to it, the OP has stacked the odds against himself. Particularly if he keeps asking the same questions for all sorts of code constructs.

Yes, there are advantages of some techniques over others. But the differences in efficiency are trade-offs, not absolutes. And the effort of premature optimisation exceeds the benefits in most cases.