Recipes 6.1, 6.2, and 6.3 explain in detail how to match integer numbers. These recipes do not allow underscores in the numbers. Their regular expressions can easily use ‹[0-9]+›, ‹[0-9A-F]+›, and ‹[01]+› to match decimal, hexadecimal, and binary numbers.

If we wanted to allow underscores anywhere, we could just add the underscore to these three character classes. But we do not want to allow underscores at the start or the end. The first and last characters in the number must be a digit. You might think of ‹[0-9][0-9_]+[0-9]› as an easy solution. But this fails to match single digit numbers. So we need a slightly more complex solution.

Our solution ‹[0-9]+(_+[0-9]+)*› uses ‹[0-9]+› to match the initial digit or digits as before. We add ‹(_+[0-9]+)*› to allow the digits to be followed by one or more underscores, as long as those underscores are followed by more digits. ‹_+› allows any number of sequential underscores. ‹[0-9]+› allows any number of digits after the underscores. We put those two inside a group that we repeat zero or more times with a asterisk. This allows any number of nonsequential underscores with digits in between them and after them, while also allowing numbers with no underscores at all.

Techniques used in the regular expressions in this recipe are discussed in Chapter 2. Recipe 2.3 explains character classes. Recipe 2.5 explains anchors. Recipe 2.6 explains word boundaries. Recipe 2.8 explains alternation. Recipe 2.9 explains grouping. Recipe 2.12 explains repetition.