Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!aioe.org!eternal-september.org!feeder.eternal-september.org!.POSTED!not-for-mail
From: Tim Watts <tw@dionic.net>
Newsgroups: comp.os.linux.misc
Subject: Re: Re (2): Linux utility with reverse index facility?
Followup-To: comp.os.linux.misc
Date: Mon, 16 May 2011 22:32:54 +0100
Organization: A noiseless patient Spider
Lines: 42
Message-ID: <6ph8a8-blt.ln1@squidward.dionic.net>
References: <iqr44b$oug$1@dont-email.me> <iqru0q$593$1@dont-email.me>
Mime-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7Bit
Injection-Info: mx04.eternal-september.org; posting-host="6oIlEBqCjOm0MjsSUEk5CA"; logging-data="22902"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX18U3HjlXPwmun2NZgvvRlLrEK45dePkiXs="
User-Agent: KNode/4.4.10
Cancel-Lock: sha1:PwoAXWhTQCoFL+NBsnFUip3SSvQ=
Xref: x330-a1.tempe.blueboxinc.net comp.os.linux.misc:1116

no.top.post@gmail.com wrote:

> What I want is the KNOWLEDGE, whether *nix designers included
> <reverse indexing> which one would use eg. to find:
> "which token/field is 'mat' in the record:
> "the cat sat on the mat".

Well, if I'm understanding this, then in perl:

#!/usr/bin/perl
$a="the cat sat on the mat"; # Input
@b=split /\s+/, $a;          # space(s) as delimeter, make array of tokens
%c=map {$_ => $i++} @b;     # Make an asociative array (tokenname => index)
print $c{'mat'}'             # try one

index is 0 based, ie 0 is the 1st token

Of course - there is a flaw in that program - hint: what happens if you 
search for the token "the" <- You didn't define what you wanted: a list of 
indices; the first one; the last one?

if the above is too long, you can abbreviate it:

perl -e '$a="the cat sat on the mat";$b={map {$_ => $i++} split /\s+/, $a}-
>{mat};print $b'

(That's one line)


This is why I use perl. Don't ask me about awk, sed, bash - I've found it 
more productive to learn one tool that ha enough functionality to do pretty 
much everything in one place - OK, not so clever on embedded systems where 
perl may not be present, but there you go...

One thing to note: the first part of split is a regex so you can get 
arbitrarily fancy on how you tokenise the string. an Unicode - no problem...

Cheers

Tim

-- 
Tim Watts