Hi all, this is my first ever post at PerlMonks.
Where I work, we need to know that "240n" humanly equates to "0.24u" == .24e-6 == 240 * 1e-9
Numerifying an input string (usually human-entered) allows for practical numerical operations/comparisons. Obviously, I made some assumptions on how a number is to be represented. For instance, "." alone is an invalid number and does not mean "0.0".
Suggestions for optimizations, a better way of thinking/writing it, using modules instead, etc. are appreciated.
#!/usr/bin/perl -wl
use strict;
#######################
## string numerifier ##
#######################
# SI prefix to exponent value conversion table
my %SI = (
Y => +24, # yotta
Z => +21, # zetta
E => +18, # exa
P => +15, # peta
T => +12, # terra
G => +9, # giga
M => +6, # mega
k => +3, # kilo
K => +3, # kilo
"" => 0, # default is no SI prefix
m => -3, # milli
u => -6, # micro
n => -9, # nano
p => -12, # pico
f => -15, # femto
a => -18, # atto
z => -21, # zepto
y => -24, # yocto
);
sub num
{
my ($numstr) = @_;
# string must contain digit
if ($numstr =~ /\d/)
{
# clean string for best chance at numerification
my ($sign, $exponent, $siprefix) = ("+", 0, "");
if ($numstr =~ /^[^.]*\d/) # does first digit appear before po
+tential decimal point?
{
$numstr =~ s/^.*?([+-]?)(\d+[.]?\d*)(?:[eE]([+-]?\d+))?([Y
+ZEPTGMkKmunpfazy])?.*/$2/;
$sign = $1 if $1;
$exponent = $3 if $3;
$siprefix = $4 if $4;
}
else
{
$numstr =~ s/^.*?([+-]?)(\d*[.]?\d+)(?:[eE]([+-]?\d+))?([Y
+ZEPTGMkKmunpfazy])?.*/$2/;
$sign = $1 if $1;
$exponent = $3 if $3;
$siprefix = $4 if $4;
}
# convert SI prefixes -- example: ".16E2m" => ".16e-1" => 0.01
+6
if ($exponent || $siprefix)
{
$exponent += $SI{$siprefix};
$numstr .= "e".$exponent;
}
# numerify string according to nearest sign
{
no warnings 'numeric';
$numstr = -+-$numstr; # high-precedence nu
+merifier
$numstr = -$numstr if $sign eq "-"; # negative if last s
+ign is "-"
}
}
else
{
$numstr = undef;
}
return($numstr);
}
##############
# test cases #
##############
print "|".num("+-- 123.mA")."|"; # output: "|0.123|"
print "|".num(" -+--.15e4u99..99.0")."|"; # output: "|-0.0015|"
print "|".num(" -+-- 789. G5e2")."|"; # output: "|789|"
print "|".(1e-6 == num("1e-6"))."|"; # output: "|1|"
print "|".(1e-6 == num("1e3n"))."|"; # output: "|1|"
print "|".(1e-6)."|".(1e3*1e-9)."|"; # output: "|1e-06|1e-06|"
print "|".(1e-6 == 1e3*1e-9)."|"; # output: "||" -- low-level FP l
+imitation, perhaps with C's strtod()
