in reply to Perl custom sort for Portuguese Lanaguage

use warnings;
use 5.016;
use utf8;
use open qw/:std :utf8/;
use Text::CSV qw/csv/; # also install Text::CSV_XS for speed
use Unicode::Collate;

my $Collator = Unicode::Collate->new(
    preprocess => sub { $_[0][0] =~ s/^(?:o|a)\s+//ir } );
my $rows = csv( in=>*DATA, sep=>"|", esc=>"\\", auto_diag=>2 );
my @sorted = $Collator->sort(@$rows);
csv( in=>\@sorted,out=>*STDOUT, sep=>"|", esc=>"\\", quote_space=>0 );

__DATA__
a cerveja|beer
o ano|year
a laranja|orange
beber|to drink
água|water
o copo de vinho|glass of wine
o copo|glass or cup
o sumo|juice

[download]

Output:

água|water
o ano|year
beber|to drink
a cerveja|beer
o copo|glass or cup
o copo de vinho|glass of wine
a laranja|orange
o sumo|juice

[download]

Update: Realized I could use $Collator->sort instead of $Collator->cmp.

Replies are listed 'Best First'.
Re^2: Perl custom sort for Portuguese Lanaguage
by ikegami (Patriarch) on Jun 14, 2020 at 00:03 UTC

    See also Unicode::Collate::Locale in the same distro. Using locale => "pt" would use Portuguese-specific rules (in any exist?) rather than a generic algorithm.

[reply]
[d/l]
Re^2: Perl custom sort for Portuguese Lanaguage
by Galdor (Sexton) on Jul 08, 2020 at 06:03 UTC
    excellent! Thank you for that! this is really useful one more thing - how can I make it ignore all lines starting with a hash - e.g. '/^#/' - is that possible with this solution? Thanks for pointers!
[reply]
      how can I make it ignore all lines starting with a hash - e.g. '/^#/' - is that possible with this solution?

      This depends on what you mean by "ignore" - I have a suspicion that perhaps your input file is in sections separated by comments, and you want the sections to be sorted individually? Could you show some short but representative sample input and the expected output for that input?

[reply]
        sure. No there is no need to sort sections individually - merely strip all blank lines and all lines start with a hash - here are a few small samples (each is a separate *.dict file): 
        # Drink
a cerveja|beer
a laranja|orange
a água|water
beber|to drink
o copo de vinho|glass of wine
o copo|glass or cup
o sumo|juice

        [download]
        and ... 
        # numbers

zero|zero
um|one
dois|two
três|three
quatro|four
cinco|five
seis|six
sete|seven
oito|eight
nove|nine
dez|ten

## 11 - 19
onze|eleven
doze|twelve
treze|thirteen
catorze|fourteen
quinze|fifteen
dezasseis|sixteen
dezasssete|seventeen
dezoito|eighteen
dezanove|nineteen

        [download]
        and even ... 
        # time
# DO NOT SORT!

# the time
o segundo|second
o minuto|minute
a ora|hour


# the day
o dia|day
a noite|night

a madrugada|early morning
a manhã|morning
a tarde|afternoon
a noite|night

o meio dia|midday
a meia noite|midnight


# days of week
a semana|week
o fim-de-semana (os fims-de-samana)|week-end

a Segunda-feira|Monday
a Terça-feira|Tuesday
a Quarta-feira|Wednesaday
a Quita-feira|Thursday
a Sexta-feira|Friday
o Sábado|Saturday
o Domingo|Sunday

# Months of the year
o mês (os meses)|month
o ano|year

Janeiro|January
Fevereiro|February
Março|March
Abril|April
Maio|May

        [download]
        so in the end they will all be "filtered" into one big dictionary output file - one per line in "pt dictionary" aphabetical order. They are kinda in markdown format so I could print each out individually without doing any sorting - but also enjoy benefit of having a "big database" of words to do word-tests, a personal dictionary, and cool stuff like that... If I want "sub-section" sorting as you suggest I will just break them out into separate files (I guess) ... Thanks !
[reply]
[d/l]
[select]

In Section Seekers of Perl Wisdom