Question

I have a string like that :

group1
  Members: 
  m/a
  m/b
group2
  Members: 
  m/c
  m/d
  m/e
group3
  No Members

I want to have a scan result like :

[["group1","a","b"],["group2","c","d","e"],["group3"]]

But I just can have :

[["group1","a"],["group2","c"],["group3", nil]]

with this regexp :

text.scan(/([^\r\n]+)\r?\n[\s\t]*(?:No |)Members[\s:]*\r?\n(?:[\t\s]*m\/(\w+)+\r?\n)*/m)

Can I do what I want only with regexp ?

Answer 1

While it’s possible to do it in a regex, it gets unwieldy, so I’d do it this way:

data = <<EOT
group1
  Members: 
  m/a
  m/b
group2
  Members: 
  m/c
  m/d
  m/e
group3
  No Members
EOT

pp data.lines.slice_before(/^group/).to_a
=> [["group1\n", "  Members: \n", "  m/a\n", "  m/b\n"],
    ["group2\n", "  Members: \n", "  m/c\n", "  m/d\n", "  m/e\n"],
    ["group3\n", "  No Members\n"]]

Cleaning up the rest to meet the question’s requirements makes it:

data.gsub(%r{\bm/}, '').split(/\n\s*/).reject{ |s| s[/\bMembers\b/] }.slice_before(/^group/).to_a
=> [["group1", "a", "b"], ["group2", "c", "d", "e"], ["group3"]]

The gist of the parsing is really in slice_before. Everything else is creating the array and cleanup.

Breaking it down:

• gsub(%r{\bm/}, '') strips the undesired m/.
• split(/\n\s*/) splits the string on line-ends into an array, while simultaneously removing the leading white-space.
• reject{ |s| s[/\bMembers\b/] } rejects any lines containing ‘Members’ as a separate word.
• slice_before(/^group/) breaks up the array into chunks starting with ‘group’ at the beginning of the string.
• to_a converts it all into an array again.