Simple use of boost::regex match groups

Question

I want to use boost::regex to change the format of a bunch of dates on the format 19991231235959 to this format 1999-12-31_23:59:59 like this:

YYYYMMDDhhmmss --> YYYY-MM-DD_hh:mm:ss
19991231235959 --> 1999-12-31_23:59:59

I use this

std::string input = "19991231235959";
boost::regex regex("^([0-9]{4})([0-9]{2})([0-9]{2})([0-9]{2})([0-9]{2})([0-9]{2})");
std::string format = "\\1-\\2-\\3_\\4:\\5:\\6";
std::string output = boost::regex_replace(input, regex, format);

which works but is there a way to get rid of the repetitions ([0-9]{2}) in regexconstruction and keep the match groups?

Also, if you set `regex_constants::escape_in_lists`, you can use `\\d` instead of `[0-9]`. — Mad Physicist, Jul 22 '16 at 14:54
Maybe regex is not the best solution; I guess it is a bit of an exercise. The reason I considered regex in the first place is that the `regex` and `format` strings can be read from configuration. — Svaberg, Jul 22 '16 at 14:56

score 0 · Answer 1 · answered Jul 22 '16 at 13:36

0

It depends on your input data. If you have a 14 digit timestamp at the beginning of every input line you could use:

boost::regex regex("^(....)(..)(..)(..)(..)(..)");

answered Jul 22 '16 at 13:36

Alex H

190
1
9

This is certainly cuter but there is still the repetition `(..)(..)(..)(..)(..)` I am thinking about something like this `(....)(..){5}` – Svaberg Jul 22 '16 at 14:46
I.e. for numbers only `^([0-9]{4})([0-9]{2}){5}` – Svaberg Jul 22 '16 at 14:47

score 0 · Answer 2 · answered Sep 22 '16 at 19:58

I am doing something similar. One thing I like to do is name my capture groups though.

boost::regex regex("\\D*(?<YYYY>\\d{4})\\D*(?<MM>\\d{2})\\D*(?<DD>\\d{2})\\D*(?<hh>\d{2})\\D*(?<mm>\\d{2})\\D*(?<ss>\\d{2})");

That gives each group the same name as your format string. So you can do things like:

int year = lexical_cast<int>(what["YYYY"]);
int month = lexical_cast<int>(what["MM"]);

Using the built in character classes also helps. The above regex will match every one of these inputs:

20160922015227
2016 09 22 01 52 27
2016_09_22-01:52:27

Simple use of boost::regex match groups

2 Answers2