perl regex to convert currency - regex

Question

i need some help in text cleaning/normalization process
i struck at a place where i need to convert a currency format
input: $100 million output: 100 million dollar
input: eur20 million output: 20 million euros
i'm using perl regex for the cleaning process, help will be appreciated if someone can help me in providing a regex to convert input to output
this is my code so far
s/([\$])([0-9\.])([million])/ $2 $3 dollars/g;
example number is $4.2 million
this is what i tried for converting dollars symbol into word "dollars" and shift it to end of phrase, but it is not providing the result as expected, it provide me ".2 million" as output

Answer 1

[...] in a regex introduces a character class, so [million] is the same as [nolim], and it matches one of those characters.
I'd create a translation table for the currencies in a hash. From the keys of the hash, you can build a regex that matches them, and use it in the replacement:
#!/usr/bin/perl
use warnings;
use strict;
use utf8;
use feature qw{ say };
my %currency = ( '$' => 'dollar', # or dollars?
eur => 'euros',
'€' => 'euros',
);
my $regex = join '|', map quotemeta, keys %currency;
for my $input ('$100 million', 'eur20 million', '€13.2 thousand') {
( my $output = $input )
=~ s/($regex)([0-9.]+ (?:million|thousand))/$2 $currency{$1}/g;
say $output;
}

Answer 2

Your regex does not give the result you claim it does.
s/([\$])([0-9.])([million])/ $2 $3 dollars/g;
With the help of the /x modifier we can add whitespace (even newlines and comments) to the pattern to improve readability. Your pattern can then be re-written as
s/([\$]) # match a literal $ and capture that as $1
([0-9.]) # match ONE digit or a dot and capture as $2
([million]) # match ONE character of 'm', 'i', 'l', 'o', 'n'
# and capture as $3
/ $2 $3 dollars/gx;
There is no way $100 million matches this pattern and results in .2 million. Possible inputs would be
$3i, $.o or $9m. They would give 3 i dollars, . o dollars, and 9 m dollars.
What you are looking for is a pattern like this:
s/\$ # a literal '$'
([\d.]+) # one or more digits or dots, like e.g. '99.5',
# captured as $1
\s+ # one or more whitespace
(million) # the literal text 'million', captured as $2
/$1 $2 dollars/gx;
(or, as a one-liner: s/\$([\d.]+)\s+(million)/$1 $2 dollars/g;)
Note that $2 in this case always is million and you could also rewrite it as s/\$([\d.]+)\s+million/$1 million dollars/g; (omitting the () around million).