From: rio4ruby <Christopher.Kleckner@...>
Date: 2007-08-17T10:50:03+09:00
Subject: Re: grep a csv?

On Aug 16, 1:43 pm, Michael Linfield <globyy3...@hotmail.com> wrote:
> Pe�a, Botp wrote:
> > From: rio4ruby [mailto:Christopher.Kleck...@gmail.com]
> > #   require 'rio'
> > #   rio('filename.csv').chomp.lines(/Blah[^,]*/) do |line,m|
> > #     rio(m) + '.csv' << line + $/
> > #   end
>
> > simply amazing. btw, how does rio handle big files, does it load them
> > whole in memory?
>
> > thanks for rio.
> > kind regards -botp
>
> it seems things have been amped a few levels of complication since my
> first few post lol. The quote above might seem like the cleanest way to
> do this, however if i use this method...ill still need the commas,
> because when u take a csv and put it in simple text, the commas are what
> seperate the columns. so maybe it should look something like this?
>
> require 'rio'
>
>  rio('filename.csv').chomp.lines(/Blah1/) do |line,m|
>    rio(m) + '.csv' << line + $/
>  end
> ###
>
> --
> Posted viahttp://www.ruby-forum.com/.

You are right, that was a poorly thought out regular expression.
One could also use Rio's csv mode (which uses the stdlib csv):

  rio('filename.csv').chomp.csv.records(/Blah\w*/) do |rec,m|
    rio(m.to_s + '.csv').csv << rec
  end

But this also is definately NOT a robust solution.