Columbia Kermit/kermit.zip/newsgroups/misc.19981211-19990422/[e-mail] _Mon Mar 22 11:16:00 1999.msg

Internet Message Format | 2020-01-01 | 6KB

Return-Path: <[e-mail]>
Received: from newsmaster.cc.columbia.edu (newsmaster.cc.columbia.edu [128.59.59.30])
    by watsun.cc.columbia.edu (8.8.5/8.8.5) with ESMTP id LAA14226
    for <[e-mail]>; Mon, 22 Mar 1999 11:16:00 -0500 (EST)
Received: (from news@localhost)
    by newsmaster.cc.columbia.edu (8.8.5/8.8.5) id LAA02385
    for [e-mail]; Mon, 22 Mar 1999 11:10:28 -0500 (EST)
X-Authentication-Warning: newsmaster.cc.columbia.edu: news set sender to <news> using -f
From: [e-mail]
Subject: Re: Associative Array in Kermit 95
Date: Mon, 22 Mar 1999 16:02:44 GMT
Organization: Deja News - The Leader in Internet Discussion
Message-ID: <7d5pit$on3$[e-mail]>
To: [e-mail]

In article <[e-mail]>,
  Mark Sapiro <[e-mail]> wrote:
> Frank da Cruz wrote:
> >>
> In article <7cttsc$47f$[e-mail]>,  <[e-mail]> wrote:
>
> : Associative array is a very useful feature in scripting language, it enables
> : Perl, Tcl etc. the implementation of comlex data structures.  The following
> : script demonstrates that associative array can also be crafted in Kermit
> : 95. The script counts the unique words in a regular english, french, german,
> : etc. text file.
> :
> : open read testfile.txt
> : if fail end 1 Can't open testfile.txt
> : assign \%n 0                                  ; init register
> : while true {
> :     read \%l                                  ; read each line
> :     if fail break                             ; until the end of file
> :     while > \flength(\%l) 0 {
> :         assign \%w \fbreak(\%l,{ })           ; split on space
> :         xif defined \m(\%w) {                 ; word already seen?
> :             _assign \%w \feval(\m(\%w) + 1)   ; incr count this word
> :         } else {
> :             _assign \%w 1                     ; init count this word
> :             increment \%n                     ; next register
> :             _assign \%n \%w                   ; register this word
> :         }
> :         assign \%l \fltrim(\fright(\%l,-
> :         \feval(\flength(\%l)-\flength(\%w)))) ; shift to next word
> :     }
> : }
> : for \%k 1 \%n 1 {
> :       assign \%w \m(\%k)                      ; get word from register
> :       echo <\m(\%w)> \%w                      ; display occurences
> : }
> :
> : This approach avoids the use of array which has to be declared in advance.
> : The script does not take into account the non alphanumeric characters.
> :
> : Dat Nguyen
> : Airline Telecommunications and Information Services
> : 770 Sherbrooke West
> : Montreal, Quebec
> : Canada H3A 1G1
> : Email dat.nguyen&sita.int
> :
> Excellent!  I've had associative arrays on my list for quite a while, but
> the list so long and time so short.  I've reformatted your script to fit
> in 80 columns.
>
> We plan to add a script library to the Kermit website -- this one will
> certainly go into it.  Other submissions are welcome too; send them in!
> (Be sure to document the Kermit program and version and other relevant
> info.)
>
> A quick glance shows this script doesn't use any new (post-C-Kermit-6.0)
> features, some of which would make it simpler and faster, for example
> the new \fword() and \fsplit() functions for extracting words from strings,
> with specified break masks (e.g. to make sure punctuation does not count
> as part of word (e.g. "thing" and "thing.").  Also you can convert each
> word to lowercase with \flower() so "Thing", "thing", and "THING" count as
> the same word, etc.
>
> Readers should take special note of the "_assign" verb, which is subtly
> different from "assign" (see p.457 of the manual).
>
> - Frank
> >>
>
> There is a problem with the above approach to implementing associative
> arrays.  Namely, if one of the 'words' used as an index happens to be
> equal to or an initial substring of a pre-defined macro (e.g. "cautious" or
> "robust") or one of the macros that gets defined implicitly when certain
> programming constructs are used (e.g. "break " or "continue"), the
> "xif defined" test gives the "wrong" result and the word doesn't get
> counted (although the macro 'word' does get (re-)defined, most likely
> with an error message from \feval()).
>
> This problem could probably be avoided by prepending some prefix to
> each word before using it as a macro name.
>
> --
> Mark Sapiro <[e-mail]>       The highway is for gamblers,
> San Francisco Bay Area, California    better use your sense - B. Dylan
>

That's right, and I was about to improve the script as follows (but it still
won't eliminate the collision 100%):

open read testfile.txt
if fail end 1 Can't not open testfile.txt
assign \%n 0
while true {
    read \%l
    if fail break
    echo \%l
    while > \flength(\%l) 0 {
        assign \%w \fbreak(\%l,{ })
        xif defined \m(word_count[\%w]) {
            _assign word_count[\%w] \feval(\m(word_count[\%w])+1)
        } else {
            _assign word_count[\%w] 1
            increment \%n
            _assign word_reg[\%n] \%w
        }
        assign \%l \fltrim(\fright(\%l,-
        \feval(\flength(\%l)-\flength(\%w))))
    }
}

for \%k 1 \%n 1 {
    assign \%w \m(word_reg[\%k])
    echo <\m(word_count[\%w])> \%w
}

Dat Nguyen
Airline Telecommunications and Information Services
770 Sherbrooke West
Montreal, Quebec
Canada H3A 1G1
Email dat.nguyen&sita.int

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own