Checking if String.scan does not match

Hi everyone,

I'm extracting 3 words from a string delimited with colons using the following approach:

  str = "test:string:here"
  str.scan(/(.+):(.+):(.+)/) do |a, b, c|
    puts a, b, c
  end

This works fine if all 3 words are present but if you pass a string with only 2 words, e.g. "test:string", the block is not executed.

I need to report an error if the match fails. This could be done like this:

  str = "test:string"
  matched = false
  str.scan(/(.+):(.+):(.+)/) do |a, b, c|
    puts a, b, c
    matched = true
  end
  if not matched
    puts "invalid format"
  end

but this is messy. Is there a neater, more ruby-like, approach to this?

Any help is greatly appreciated,
Alex

Alex Allmont wrote:

I'm extracting 3 words from a string delimited with colons using the
following approach:

  str = "test:string:here"
  str.scan(/(.+):(.+):(.+)/) do |a, b, c|
    puts a, b, c
  end

This works fine if all 3 words are present but if you pass a string with
only 2 words, e.g. "test:string", the block is not executed.

I need to report an error if the match fails.

scan is the wrong tool here, because it is intended to match the pattern
multiple times in the source string. I think all you want is a basic
regexp match:

  str = "test:string:here"
  if /\A(.+):(.+):(.+)\z/ =~ str
    puts $1, $2, $3
  else
    puts "invalid format"
  end

Alternatively written as

  case str
  when /\A(.+):(.+):(.+)\z/
    puts $1, $2, $3
  else
    puts "invalid format"
  end

which is useful when there are several different patterns to match
against the same string.

Beware lines which contain more than 2 colons, because . matches colons
as well as non-colons. For proper validation you probably want

  /\A([^:]+):([^:]+):([^:]+)\z/

Note that . doesn't match newline. If str may legitimately end with \n,
but you don't want to capture the \n in $3, then use \Z instead of \z.

If you are 100% positive that str doesn't contain more than one line
then you can use ^ and $ instead of \A and \Z

···

--
Posted via http://www.ruby-forum.com/.