Regex for striping non alphabetic and non numeric characters - ruby-on-rails

Question

Total programming newbie here. In ruby, how would I go about striping the following string of non alphabetic and non numeric characters and then split the string into an array by splitting it through spaces.
Example
string = "Honey - a sweet, sticky, yellow fluid made by bees and other insects from nectar collected from flowers."
Into this
tokenized_string = ["Honey", "a", "sweet", "sticky", "yellow", "fluid", "made", "by", "bees", "and", "other", "insects", "from", "nectar", "collected", "from", "flowers"]
Any help would be much appreciated!

Answer 1

I'd use:
string = "Honey - a sweet, sticky, yellow fluid made by bees and other insects from nectar collected from flowers."
string.delete('^A-Za-z0-9 ').split
# => ["Honey",
# "a",
# "sweet",
# "sticky",
# "yellow",
# "fluid",
# "made",
# "by",
# "bees",
# "and",
# "other",
# "insects",
# "from",
# "nectar",
# "collected",
# "from",
# "flowers"]
If you're trying to remove everything but alphanumerics, then the \w character class can't be used because it is defined as [A-Za-z0-9_], which allows _ to leak in or squeeze through. Here's an example:
'foo_BAR12'[/\w+/] # => "foo_BAR12"
That matched the entire string, including _.
'foo_BAR12'[/[A-Za-z0-9]+/] # => "foo"
That stopped at _, because the class [A-Za-z0-9] doesn't include it.
\w should be considered a matching pattern for variable names, NOT for alphanumerics. If you want a character class for alphanumerics, look at the POSIX \[\[:alnum:\]\] class:
'foo_BAR12'[/[[:alnum:]]+/] # => "foo"

Answer 2

There are a lot of possibilities, e.g.:
string.gsub(/\W/) { |m| m if m == ' ' }.split
or, even clearer:
string.gsub(/\W/) { |m| m if m.strip.empty? }.split

Answer 3

Very simple. The following gives you the array you want without your having to use split:
string.scan(/\w+/)
Play around with it on Rubular.com.

Answer 4

Do as belowe using String#scan
string = "Honey - a sweet, sticky, yellow fluid made by bees and other insects from nectar collected from flowers."
string.scan(/[a-zA-Z0-9]+/)
# => ["Honey",
# "a",
# "sweet",
# "sticky",
# "yellow",
# "fluid",
# "made",
# "by",
# "bees",
# "and",
# "other",
# "insects",
# "from",
# "nectar",
# "collected",
# "from",
# "flowers"]