Why does Ruby String#split not treat consecutive trailing delimiters as separate entities?

Question

I'm reading from a government text file in which $ is used as the delimiter, but I don't think delimiter character matters...

So this is expected:

'a$b$c$d'.split('$')
# => ["a", "b", "c", "d"]

In the datafiles I'm working with, the column headers row (the first line) are uniformly filled in, i.e. there is no empty header, as in:

'a$b$$d'
# or: 
'a$b$c$'

However, each row may have consecutive trailing delimiters such as:

"w$x$$\r\n"

Usually, I read each line and chomp it. But this causes String#split to treat the final two delimiters as one column:

"w$x$$\r\n".chomp.split('$')
# => ["w", "x"] 

Not doing the chomp gets me the desired result, though I should chomp the last element:

"w$x$$\r\n".split('$')
# => ["w", "x", "", "\r\n"]

So either I have to:

• chomp the line if the final non-newline characters are NOT consecutive delimiters
• preserve the newline, do the split, and then chomp the final element IF the final characters are consecutive delimiter

This seems really awkward...am I missing something here?

Answer 1

You need to pass a negative value as the second parameter to split. This prevents it from suppressing trailing null fields:

"w$x$$\r\n".chomp.split('$', -1)
# => ["w", "x", "", ""]

See the docs on split.