Boost::Split using whole string as delimiter

Question

I would like to know if there is a method using boost::split to split a string using whole strings as a delimiter. For example:

str = "xxaxxxxabcxxxxbxxxcxxx"

is there a method to split this string using "abc" as a a delimiter? Therefore returning:

Results would be the string "xxaxxxx" and "xxxxbxxxcxxx".

I am aware of boost::split using the "is_any_of" predicate, however invoking is_any_of("abc") would result in splitting the string at the single character 'a', 'b', and 'c' as well, which is not what I want.

You should be able to use some combination of `equals` and the string value to be used as delimiter, but the specifics elude me right now. — K-ballo, Sep 15 '11 at 20:21
possible duplicate of [Split on substring](http://stackoverflow.com/questions/3739280/split-on-substring) — Cubbi, Sep 15 '11 at 21:09

Mythli · Answer 1 · 2011-09-15T20:49:41.947

8

Yes there is a way (this is a way I know, maybe there is a better way) Use boost::algorithm::split_regex to split character sequences where delimiters are regular expressions.

Example:

vector< string > result;
boost::algorithm::split_regex( result, str, regex( "^((?!abc)*abc(?!abc)*)*$" ) ) ;
copy( result.begin(), result.end(), ostream_iterator<string>( cout, "\n" ) ) ;

edited Sep 15 '11 at 20:49

answered Sep 15 '11 at 20:30

Mythli

5,995
2
24
31

1

Thanks, this solution works, however it requires boost to be built. Currently I am using header files only. – andre Sep 16 '11 at 15:05
2

#include – D. A. Aug 29 '14 at 17:52

score 6 · Accepted Answer · answered Sep 15 '11 at 20:43

6

split_regex as suggested by @Mythli is fine. If you don't want to deal with regex, you can use ifind_all algo, as is shown in this example. You receive iterator_range (begin/end) of all occurrences of you delimiter. Your tokens are between them (and at the beginning and end of string).

answered Sep 15 '11 at 20:43

Andriy Tylychko

15,967
6
64
112

Thanks, this solution fit my needs perfectly. – andre Sep 16 '11 at 15:05

score 0 · Answer 3 · answered Dec 27 '20 at 09:31

0

If you know what characters your input string is comprised of (e.g. a simple English sentence, with only EN characters), you can surround your word_delimiter with a special_symbol, replace all appearances of your word delimiter with it and split by the special_symbol.

For example, I used '%' here:

std::vector<std::string> sentence_parts;
boost::replace_all(sentence, word, "%" + word_delimiter + "%");
boost::split(sentence_parts, sentence, boost::is_any_of("%"));

You need to be sure that your input string will never have the special symbol in it, otherwise it won't work well!

answered Dec 27 '20 at 09:31

may21

31
1
3

Why would this be preferable to the more straightforward approach Mythli already described? – Sneftel Dec 27 '20 at 09:32
It's not necessarily preferable, it's another way to do it. Personally, I prefer not to deal with regexes... – may21 Dec 27 '20 at 09:40

Boost::Split using whole string as delimiter

3 Answers3