From: Brian Candler <B.Candler@...>
Date: 2007-02-07T18:41:11+09:00
Subject: Re: Parse csv similar file

On Wed, Feb 07, 2007 at 05:47:26PM +0900, Rebhan, Gilbert wrote:
> i get a folder
> 
> /timestamp
>   metafile.txt
>    /INCLIB
>    /PLI
> 
> metafile looks like that =
> APLVZDT;INCLIB;E050441;AP013;240107;0730
> AP400ER;INCLIB;E023889;AP013;240107;0730
> AP540RBP;INCLIB;E052337;AP013;240107;0730
> AP700PA;INCLIB;E050441;AP013;240107;0730
> ... more lines
> 
> field 1 is a filename
> field 2 is a foldername, shows whether path is /INCLIB/file or /PLI/file
> field 3 is a ticketnr
> field 4 is a username
> field 5 is a date
> field 6 is a timestamp
> 
> 
> i need to parse the metafile and =
> 
> 1. create a folderstructure for every ticketnr that occurs, f.e.
> 
> /E050441
>    /INCLIB
>    /PLI
> 
> and put all the files that belong to that ticket
> (means the line with the filename contains that ticketnr)
> in the subfolder which is field 2
> 
> 2. create a file in the root of the /ticketnr folder
>     which contains the rest of a dataset (line), means =
> 
> field 4
> field 5
> field 6
> 
> which are the same for every file with the same ticketnr
> 
> the format might look like
> 
> user=...
> date=...
> time=...
> 
> have to decide it later.
> 
> 
> I thought with =
> 
> File.open("mycsvfile", "r").each do |line|
>         if line =~ /(\w+.?);(\w+);(\w+);(\w+);(\w+);(\w+)/
>          
>          efas<<$3.to_s<<',' unless efas.include?($3.to_s)
> 
> i get an array with all ticketnr
> then i create a folderstructure for every index in that array
> and put the files in it, but i don't get it.
> 
> Any ideas ?

I'd do all the work on-the-fly. Untested code:

require 'fileutils'
SRCDIR="/path_to_src"
DSTDIR="/path_to_dst"

def copy_ticket(filename, folder, ticket, user, date, time)
  srcdir = SRCDIR + File::SEPARATOR + folder
  dstdir = DSTDIR + File::SEPARATOR + ticket + File::SEPARATOR + folder
  FileUtils.mkdir_p(dstdir)
  FileUtils.cp(srcdir + File::SEPARATOR + filename,
               dstdir + File::SEPARATOR + filename)

  # write out status file
  statusfile = dstdir + File::SEPARATOR + "status.txt"
  unless FileTest.exists?(statusfile)
    File.open(statusfile, "w") do |sf|
      sf.puts "user=#{user}"
      sf.puts "date=#{date}"
      sf.puts "time=#{time}"
    end
  end
end

def process_meta(f)
  f.each_line do |line|
    next unless line =~ /^(\w+);(\w+);(\w+);(\w+);(\w+);(\w+)$/
    copy_ticket($1,$2,$3,$4,$5,$6)
  end
end

# Main program
File.open("mycsvfile") do |f|
  process_meta(f)
end

If you want to build up a hash of ticket IDs seen, you can do that in
process_meta as well. I'd pass in an empty hash, and update it in the
each_line loop.

HTH,

Brian.