Trimming a vector of strings - c++

Question

I have an std::vector of std::strings containing data similar to this:
[0] = ""
[1] = "Abc"
[2] = "Def"
[3] = ""
[4] = "Ghi"
[5] = ""
[6] = ""
How can I get a vector containing the 4 strings from 1 to 4? (i.e. I want to trim all blank strings from the start and end of the vector):
[0] = "Abc"
[1] = "Def"
[2] = ""
[3] = "Ghi"
Currently, I am using a forward iterator to make my way up to "Abc" and a reverse iterator to make my way back to "Ghi", and then constructing a new vector using those iterators. This method works, but I want to know if there is an easier way to trim these elements.
P.S. I'm a C++ noob.
Edit
Also, I should mention that the vector may be composed entirely of blank strings, in which case a 0-sized vector would be the desired result.

Answer 1

How about this, with a predicate:
class StringNotEmpty
{
bool operator()(const std::string& s) { return !s.empty(); }
};
Now to trim:
vec.erase(std::find_if(vec.rbegin(), vec.rend(), StringNotEmpty()).base(), vec.end());
vec.erase(vec.begin(), std::find_if(vec.begin(), vec.end(), StringNotEmpty()));
There might be an off-by-one on the .base() call, but the general idea should work.

Answer 2

Your approach is reasonable. Another approach is to find the first string, and then copy all successive strings to the beginning (and following) elements in the vector, after which you trim the end of the vector.
Unless this is a critical piece of code that is slowing down your application, it matters more that it works than that you have the most efficient implementation possible.

Answer 3

What you've done is good. But if you want to shorten a single container "in place" rather than make a copy of a subrange, you can erase the other ranges from the vector.
// Probably similar to what you already have: Find iterators for the range to keep.
std::vector<std::string>::iterator start=strs.begin(), stop=strs.end();
start = strs.begin();
while (start != stop && start->empty()) ++start;
if (start == stop) {
// all strings were empty!
strs.clear();
} else {
while (start != --stop && stop->empty()) ;
++stop;
// std::vector<std::string>(start, stop) is the desired subrange
strs.erase(stop, strs.end());
strs.erase(strs.begin(), start);
}
But I agree with #Nathan: If you what you already have makes more sense to you than this, keep it, unless you know or strongly suspect that there will be an awful lot of strings involved.

Answer 4

typedef std::vector<std::strings> Strings;
Strings myStrings;
//... currently myStrings contains "","Abc","Def","","Ghi","",""
Strings::iterator itNewBegin = myStrings.begin();
Strings::iterator itNewEnd = myStrings.end();
std::advance(itNewBegin,1);
std::advance(itNewEnd,4);
String myTrimmedStrings(itNewBegin,itNewEnd);
//... currently myTrimmedStringscontains "Abc","Def","","Ghi"
By curiosity, I would like to see your code with the use of the reverse iterator.
I don't grasp the way you contruct new vector from two iterators that have different directions.

Answer 5

Iterators are definitely the way to go. They make the code much more readable and intuitive.
#include <algorithm>
typedef std::vector<std::string> strVec_t;
bool notEmpty(std::string s){
return !s.empty();
}
void trim(strVec_t& vec){
//get the first and last non-empty elements
strVec_t::const_iterator first = std::find_if(vec.begin(), vec.end(), notEmpty);
strVec_t::const_reverse_iterator last = std::find_if(vec.rbegin(), vec.rend(), notEmpty);
//if every element is an empty string, delete them all
if (first == vec.end()){
vec.clear();
//make a new vector from the first to the last non-empty elements
} else {
vec = strVec_t(first, last.base());
}
}