- 20.1. Overview of C++11 Regular Expressions
- 20.2. Dealing with Escape Sequences (\)
- 20.3. Constructing a RegEx String
- 20.4. Matching and Searching Functions
- 20.5. "Find All," or Iterative, Searches
- 20.6. Replacing Text
- 20.7. String Tokenizing
- 20.8. Catching RegEx Exceptions
- 20.9. Sample App: RPN Calculator
20.6. Replacing Text
One of the most powerful regular-expression capabilities is to selectively search-and-replace patterns within a string of text. Here’s one possible use (out of zillions): to transform a target string by replacing each repeated pair of words with just one word.
For example, given this text:
The cow cow jumped over the the moon.
it would be useful to produce a string consisting of:
The cow jumped over the moon.
The regex_replace function performs this task by returning the transformed string. It has the following syntax:
regex_replace(target_string, regex_obj, replacement_pattern_str);
The replacement_pattern_str is a string that can contain the following special sequences (in addition to ordinary characters).
Refers to the entire matched string.
Refers to the nth group within the matched string. For example, “$1” refers to the first group of characters tagged by the regex object; “$2” refers to the second group of tagged characters (if there is one), and so on. The example that follows should clarify.
A literal dollar sign ($).
The following declarations set up a search-and-replace designed to fix the repeated-word pattern, replacing it, where found, with one copy of the word.
using std::regex; using std::regex_replace; using std::string; regex reg1("([A-Za-z]+) \\1"); // Find double word. string replacement = "$1"; // Replace with one word.
With these objects defined, the following statements execute search-and-replace on the string shown earlier.
string target = "The cow cow jumped over the the moon."; string result = regex_replace(target, reg1, replacement); std::cout << result << std::endl;
The output is:
The cow jumped over the moon.
which is what we wanted.
Let’s review how this works. When the text “cow cow” was matched by the regular-expression object, the first occurrence of “cow” was tagged because it matched the expression inside the parentheses: “([A-Za-z]+)”. The rest of the expression, “\\1”, indicated that the regex object then needed to match a space, followed by a recurrence of the tagged characters, to match the overall expression. Therefore, “cow” gets tagged and “cow cow” matches the entire regular expression.
The replacement pattern, “$1”, causes the matched text—“cow cow”—to be replaced by “cow”, the tagged group. Suppose the replacement pattern were “XX$1YY$1ZZ$1”. Then the replacement text would have been “XXcowYYcowZZcow” and that would have replaced “cow cow”.
Characters not matched by the regex object, reg1, are just copied into the result as they are. So, for example, the words “jumped over” are copied without being transformed.
Here’s an example of another regex object and replacement-pattern string: When used with the call to regex_replace shown earlier, these result in the switching of two words separated by an ampersand (&). For example, “boy&girl” would be replaced by “girl&boy” and vice versa.
regex reg1("([A-Za-z]+)&([A-Za-z]+)"); // Find word&word string replacement = "$2&$1"; // Switch order.
The regex_replace function is particularly convenient. It isn’t necessary to iterate through the target string. Instead, regex_replace carries out replacements on all the substrings matching the pattern in the regex object, while leaving the rest of the text alone.