Consecutive/Multiple Delimiters? about strutil HOT 4 CLOSED

Corollary: Are you open to a pull request adding this feature? I'm thinking something like:

std::vector<std::string> multi_split(const std::string& str, const std::string& delims, bool collapse = true);

from strutil.

Sure, go ahead! If you write a working algorithm and test(s) for it, I'll merge your code into the code base :) Otherwise, I'll probably write my own implementation of this, but this could happen in ~3 weeks (best case scenario).

from strutil.

While implementing this, I'm noticing as few inconsistencies in the semantics of split for char and str inputs:

Empty string handling
Splitting on an empty string results in different output depending on delimiter type:

split("", ';').size()  == 0; // split return []
split("", ">=").size() == 1; // split returns [""]

I'm inclined to the second approach, as in python:

"".split(";') == [""] // True

Trailing Delimiter Handling
Splitting strings with trailing delimiters returns different results:

split("abc;", ';').size()  == 1;  // split returns ["abc"]
split("abc>=", ">=").size == 2; // split returns ["abc", ""]

Again, python favors the latter:

"abc;".split(";") == ["abc", ""] // True

Thoughts? I'm thinking I'll adjust the char-splitting version to match python semantics.

from strutil.

Sure, match char versions to string equivalents - these functions should have the same output (as their names are the same) not to mislead the end user. You can also add additional tests to tests/test_cases.cpp. I appreciate your contribution! :)

from strutil.