From: Johan Dahl <Johan.Dahl@...>
Date: 2001-05-30T06:45:21+09:00
Subject: [ruby-talk:15962] Re: .scan() using '/' as the regular expression delimiter

>err, my current code looks something like this (noting that it does not in
>fact work)
>
>regExpr1 = /\d+'/'\s+'/'\d+/
>@day, @month, @year = date.scan(regExpr1)
>
>-Luke
>
>
>"Luke Crook" <lcluke@wsbnet.com> wrote in message
>news:9f12ej$1a5@dispatch.concentric.net...
>>  I am trying to parse an Apache log file. The date field looks like:
>>      [22/May/2001:23:54:41 -0700]
>>
>>  How would I use the .scan() method when '/' is the delimiter (I'm trying
>to
>>  store day, month, year as seperate fields).
>>  I have tried \/, "/" , '/', and even [/] but these do not work.
>>
>>  -Luke Crook
>>
>>

Hi

You are using the wrong function. scan should match the whole pattern 
for each item in the produced array so for example 
"1/2/3".scan(/\d+/) ==> ["1", "2", "3"]. Use the regular expression 
matching operator =~ (equals tilde) instead.

Like this

date = "[22/May/2001:23:54:41 -0700]"

date =~ /(\d+)\/(\S+)\/(\d+)/

@day, @month, @year = $1, $2, $3

Notice \S matches any non-whitespace character. \s matches whitespace


-- 
Johan Dahl, Research Engineer
Department of Linguistics and Phonetics, Lund University
Address: Helgonabacken 12, 223 62 Lund, SWEDEN
Tel: +46-(0)46-2228443 Fax: +46-(0)46-2224210
E-mail: mailto:Johan.Dahl@ling.lu.se
WWW: http://www.ling.lu.se/persons/Johan/