#!/usr/bin/perl
use 5.014000;
use warnings;
our $VERSION = "0.01 - 20201021";
our $CMD = $0 =~ s{.*/}{}r;
sub usage {
my $err = shift and select STDERR;
say "usage: $CMD [-c #] [-r #] [-b B] file.csv";
say " -c N --columns=N Split on max N columns per file";
say " -r N --records=N Split on max N records per file";
say " -b B --base=B Use B as base for new file names";
exit $err;
} # usage
use CSV;
use Getopt::Long qw(:config bundling);
GetOptions (
"help|?" => sub { usage (0); },
"V|version" => sub { say "$CMD [$VERSION]"; exit 0; },
"c|columns=n" => \ my $opt_c,
"r|rows|records=n" => \ my $opt_r,
"b|base=s" => \ my $opt_b,
"v|verbose:1" => \(my $opt_v = 0),
) or usage (1);
my $inf = shift or usage (1);
-s $inf or die "$inf is empty";
$opt_b //= $inf =~ s{\.csv$}{}ri;
my $x = "01";
open my $ih, "<", $inf or die "$inf: !$\n";
my $csv = Text::CSV_XS->new ({ binary => 1, auto_diag => 1 });
my @h = $csv->header ($ih, { bom => 1 });
if (!$opt_c or @h <= $opt_c) {
# idx, fh, fields
@h = ([ "", undef, @h ]);
}
else {
my @c;
my $i = "A";
while (@h) {
push @c => [ $i++, undef, splice @h, 0, $opt_c ];
}
@h = @c;
}
sub next_oh {
for (@h) {
my ($idx, $fh, @fld) = @$_;
$fh and close $fh;
my $fn = $opt_b;
$opt_r and $fn .= "-$x";
$idx and $fn .= "-$idx";
$fn .= ".csv";
$fn eq $inf and die "I WILL NOT OVERWRITE $inf!";
open $fh, ">", $fn or die "$fn: $!\n";
$csv->say ($fh, \@fld);
$_->[1] = $fh;
}
$x++;
} # next_oh
$opt_r or next_oh ();
my $n = 0; # Skip header in record number
while (my $r = $csv->getline_hr ($ih)) {
if ($opt_r) { $n++ % $opt_r or next_oh () };
for (@h) {
my ($idx, $fh, @fld) = @$_;
$csv->say ($fh, [ @{$r}{@fld} ]);
}
}
close $ih;
close for grep { defined } map { $_->[1] } @h;
__END__
file.csv:
a,b,c,d,e
1,2,3,4,5
2,2,3,4,4
3,2,3,4,3
4,2,3,4,2
5,2,3,4,1
csv-split -r 3 file.csv
file-01.csv
a,b,c,d,e
1,2,3,4,5
2,2,3,4,4
3,2,3,4,3
file-02.csv
a,b,c,d,e
4,2,3,4,2
5,2,3,4,1
csv-split -c 3 file.csv
file-A.csv file-B.csv
a,b,c d,e
1,2,3 4,5
2,2,3 4,4
3,2,3 4,3
4,2,3 4,2
5,2,3 4,1
csv-split -r 3 -c 3 file.csv
file-01-A.csv file-01-B.csv
a,b,c d,e
1,2,3 4,5
2,2,3 4,4
3,2,3 4,3
file-02-A.csv file-02-B.csv
a,b,c d,e
4,2,3 4,2
5,2,3 4,1
Generated by dwww version 1.15 on Sun Jun 30 10:12:52 CEST 2024.