I have a file that contain list of numbers that lo开发者_如何学Coks like this:
10^-92
2 10^-14
10^-105
3 10^-20
To explain a bit further 10^-92
is essentially 1E-92
and 2 10^-14
is 2E^-14
.
Is there a compact way to convert the number in above file into Perl number? At the end I want to sort these numbers numerically.
If all your numbers require the same transform to become numeric to Perl, why not use a regex substitution:
use 5.010;
my @nums = ('10^-92', '2 10^-14', '10^-105', '3 10^-20');
s{ (\d*) \s* 10\^ }{ ($1 || 1) . 'e' }xe for @nums;
say "@nums"; # 1e-92 2e-14 1e-105 3e-20
If you want a Perl number, just go straight to the number without the eval. Store the parts of the number that you extract and use them in a cached-key sort:
use 5.010;
while( <DATA> )
{
chomp;
m/
(?:
(?<coefficient> \d+ (?:\.\d+)? )
\s+
)?
(?<base> (\d+) )
\^
(?<exponent> -?\d+ )
/x;
push @numbers,
[ ( $+{coefficient} // 1 ), $+{exponent}, $_ ];
}
my @sorted = sort {
$a->[1] <=> $b->[1]
||
$a->[0] <=> $b->[0]
} @numbers;
foreach my $number ( @sorted )
{
print "$number->[-1]\n";
}
__DATA__
10^-92
2 10^-14
10^-105
3 10^-20
You can boil that down to a Schwartzian Transform:
use 5.010;
print
map {
$_->[-1]
}
sort {
$a->[1] <=> $b->[1]
||
$a->[0] <=> $b->[0]
}
map {
m/
(?:(?<c> \d+ (?:\.\d+)? ) \s+ )? # coefficient
(?<b> (\d+) ) # base
\^
(?<e> -?\d+ ) # exponent
/x;
[ ( $+{c} // 1 ), $+{e}, $_ ];
} <DATA>;
__DATA__
10^-92
2 10^-14
10^-105
3 10^-20
$ perl -ne 's/ /*/;s/\^/E/;print "\n".eval $_' file
2e-13
1e-91
2e-13
1e-104
3e-19
精彩评论