Question

I have a string seperated by \t and ,, but the number of \t is not fixed, for example :

a=["seg1\tseg2\t\tseg3,seg4"]

seg2 and seg3 is seperated by two \t.

So I try to split them by

a.split(/\t+|,/)

it print the right anwser :

["seg1", "seg2", "seg3", "seg4"]

And I also try this

a.split(/[\t+,]/)

but the answer is

["seg1", "seg2", "", "seg3", "seg4"]

Why ruby print different results?

Answer 1

Because \t+ inside [] does not mean “one or more tabs”, it means “a tab or a plus”. Since it finds two consecutive tabs, it splits twice, and the string in the middle becomes empty.

Most special characters, like . + * ? etc, when placed in an interval become “regular” characters. There are some exceptions, like ^ (which negates the interval when placed at the beginning), the \ (that escapes the next character(s), just like it does outside intervals) and the ] (that closes the interval; another [ is also disallowed there). So, [\t+,] actually means '\t' or '+' or ','.

Unfortunatly, I don’t know any reference for the full set of characters that need or don’t need escaping inside an interval. In doubt, I tend to escape just to be sure. In any case, an interval will always match a single character only, if you want something different you must put your quantifier outside the interval. (For example: [\t,]+, if you also admit two commas in a row; otherwise, your first regex is really the correct one)