fun with hashes!

How about using a hash to keep track of which things you've already handled?

my %seen;
for my $value (@values) {
next if $seen{$value}++;
# Do processing with $value here.
}

--
Keith C. Ivey <keith@iveys.org>
Washington, DC

>How about using a hash to keep track of which things you've already handled
?
That's just the afore-mentioned count
--
Free map of local environmental resources: http://CambridgeMA.GreenMap.org
--
MOTD on Pungenday, the 36th of The Aftermath, in the YOLD 3173:
Wibble.

Uri wrote:
>  isa
>    is a key valid? (not same as in a set).
>    ISA usually works on fixed sets of keys

Well, in a strict sense it IS the same, as it is just one
of many possible set operations.  I suppose you could say
it's a subset of sets. :-)  I think you're trying to focus
on the difference between sets of strings, which are the
hash keys() directly, and sets of other things (e.g. objects),
which are hash values().  You also emphasize the const-ness
of the "isa" key set, which I think is rather too narrow.
I think a better approach would be to talk about sets in
general, the many set operations (including the *missing*
operations, which have to be kludged in Perl), the set-
related CPAN modules, and how all these things can be used.
JMHO.  Sorry if someone already made this point and I missed
it.

--
John Douglas Porter




 ________________________________________
____________________________________
________
Be a better pen pal.
Text or chat with friends inside Yahoo! Mail. See how.  http://overview.mail.yahoo
.com/

>>>>> "JDP" == John Douglas Porter <johndporter@yahoo.com> writes:

JDP> Uri wrote: 

JDP> Well, in a strict sense it IS the same, as it is just one
JDP> of many possible set operations.  I suppose you could say
JDP> it's a subset of sets. :-)  I think you're trying to focus
JDP> on the difference between sets of strings, which are the
JDP> hash keys() directly, and sets of other things (e.g. objects),
JDP> which are hash values().  You also emphasize the const-ness
JDP> of the "isa" key set, which I think is rather too narrow.
JDP> I think a better approach would be to talk about sets in
JDP> general, the many set operations (including the *missing*
JDP> operations, which have to be kludged in Perl), the set-
JDP> related CPAN modules, and how all these things can be used.
JDP> JMHO.  Sorry if someone already made this point and I missed
JDP> it.

i do plan on covering sets in their own right along with intersection
and union operations too (right out of the FAQ). but i still think isa
is a (sic) concept that is important enough to cover on its own. sure it
is a set but a very specific type with its own name. the names of these
concepts are important (almost like design patterns which i despise
:). i have seen too many newbies and even some mildly experienced perl
hackers don't know all these hash uses (or their names). ever see the
(virtual) glassy eyes when you mention dispatch tables in polite public?
:)

i am aiming this for my class which is a mix of experienced hackers down
to some ex-cobol coders who code perl because they get paid for it. and
it is also for future classes or writings. so i want to cover as many
uses at a basic level and not get into subtleties such as which are
really variations of each other. they are all hash variations when you
come down to that so the context and name of each use must be clearly
explained.

i will post here the url for these slides (with mistakes! :) when i get
them done. please be kind. :)

thanx,

uri

--
Uri Guttman  ------  uri@stemsystems.com  -------- [url]http://www.stemsystems.com[/url
]
--Perl Consulting, Stem Development, Systems Architecture, Design and Coding
-
Search or Offer Perl Jobs  ----------------------------  [url]http://jobs.perl.org[/url
]

Jerrad Pierce wrote: 
> That's just the afore-mentioned count

True, I guess, but lots of these uses are the same.  In Uri's original
post, "isa" could be considered the same as "sets", and "records" the
same as "data structures".

With a %seen hash, the count is being calculated, but it's not being
used except as a boolean, and it's a fairly common idiom.  I'd say
that's a separate use, but of course Uri can judge.

--
Keith C. Ivey <keith@iveys.org>
Washington, DC

Don't know if this falls into categories already
covered (set of active objects in a class? counting?),
but I often find a hash of open filehandles useful.  e.g. to
divide a big file into several smaller ones, using the
distinct values of one of the fields as filenames :

perl -MIO::File -ane '($fds{$F[1]} ||= IO::File->new(">$F[1]"))->print($_)' 
< bigfile

Brian

On Friday, November 23, Uri Guttman wrote:
> one goal i have is to
> list and explain as many different uses for hashes as reasonably
> possible.

I often use them to divide one big file into several
small files named after some field in each line, e.g.

perl -MIO::File -ane '($fds{$F[1]} ||= IO::File->new(">$F[1]"))->print($_)' 
< bigfile

Brian

>>>>> "s" == shmem  <gm@cruft.de> writes:


s> You forgot an obscure corner of hashes: hashes in scalar context.
s> Here's a a use - calculate the next power of 2 of a given number:

s> sub next_power_of_two {
s>     my %s;
s>     @s{1..shift} = ();
s>     %s =~ '/';
s>     return $';
s> }

s> ;-)

GACK!!!

and that assume knowledge of the internals of hashes. not a good thing
to teach! :)

uri

--
Uri Guttman  ------  uri@stemsystems.com  -------- [url]http://www.stemsystems.com[/url
]
--Perl Consulting, Stem Development, Systems Architecture, Design and Coding
-
Search or Offer Perl Jobs  ----------------------------  [url]http://jobs.perl.org[/url
]

>>>>> "KI" == Keith Ivey <keith@iveys.org> writes:

KI> Jerrad Pierce wrote: 

KI> True, I guess, but lots of these uses are the same.  In Uri's original
KI> post, "isa" could be considered the same as "sets", and "records" the
KI> same as "data structures".

KI> With a %seen hash, the count is being calculated, but it's not being
KI> used except as a boolean, and it's a fairly common idiom.  I'd say
KI> that's a separate use, but of course Uri can judge.

i have never run into a situation where i may process some items and
later some more and need to track them with a hash. it doesn't mean that
isn't a good use but i don't see why it couldn't be done
differently.

then again, i eschew else clauses which is rare. i just seem to code in
a way that i very rarely need them. and i am not talking obscure code,
just judicious use of statement modifiers, subs, logic order that
generally eliminates the need for else. others will use else as they
please.

thanx,

uri

--
Uri Guttman  ------  uri@stemsystems.com  -------- [url]http://www.stemsystems.com[/url
]
--Perl Consulting, Stem Development, Systems Architecture, Design and Coding
-
Search or Offer Perl Jobs  ----------------------------  [url]http://jobs.perl.org[/url
]

Uri wrote:
> i still think ISA is a (sic) concept that is important enough to
> cover on its own.  sure it is a set but a very specific type
> with its own name. the names of these concepts are important
> (almost like design patterns which i despise :).

If names are that important to you, then you should use the right
ones.  I've never seen this set membership test thing you've
illustrated called "isa".  To most programmers, "isa" means
something else.  Oops - s/else/different/; sorry. ;-)

--
jdp



 ________________________________________
____________________________________
________
Be a better pen pal.
Text or chat with friends inside Yahoo! Mail. See how.  http://overview.mail.yahoo
.com/