From: Robert Klemme <shortcutter@...>
Date: 2009-12-13T20:20:14+09:00
Subject: Re: Question about file management [read]

On 12.12.2009 21:05, Panagiotis Atmatzidis wrote:
> A salut to the list,
>
> I need read info from a file, extract the info that I want which is the date and IP and dump them into a database (mysql).
>
> However, this file gets constantly updated, sits on /var/log/.
>
> Until now I have used just the file.open and file.read methods. However, I feel that both ways are not suitable for what I want to do, because I will need to make a selection between old and new entries every time my script reads the file to extract info.
>
> What is the *standard* (or best) way to manage this situation?
>
> Here is the class which reads the file:
>
> class ReadIPs
>    def self.ip(log)
>      file = File.open(log)

Ha!  There you are not using block form of File.open and you do not have 
the #close in ensure! :-)

>      file.each_line do |line|
>        if line.include?("Ban")
>           a = line.split(" ")
>           date = a[0]
>           ip = a[6]
>           puts "date: #{date} IP: #{ip}"
>        end
>      end
>      file.close
>    end
> end
>
> I use File.open now but I should use something like File.append or similar? Note that I will never write to this file, it's just the file that contains the info. Every time a new line appears that contains the word "Ban" I want to trigger this...
>
> Best regards

I can't test it right now but this might be an option.

File.open log do |io|
   loop do
     io.each do |line|
       # whatever...
     end

     # v2: pos = io.tell
     sleep 10 # whatever interval is appropriate
     # v1: io.seek(0, IO::SEEK_SET)
     # v2: io.seek(pos, IO::SEEK_CUR)
   end
end

Maybe you have to do one of the variants to clear eof flag:

v1: seek to offset 0 of current position
v2: fetch the position before the sleep and then seek to that absolute 
position

Kind regards

	robert


-- 
remember.guy do |as, often| as.you_can - without end
http://blog.rubybestpractices.com/