While this obviously lends itself to a CSV module, wouldn't it be considerably simpler just to use split?

my @lines;

while (<FILE>) {
    my @fields = (split /,/, $line)[1,2,3,4,5,9,11,12,13,15,16,17,18,1
+9,20,21,22,23,24];
    for (0..$#fields) {
        $fields[$_] = qq("$fields[$_]") unless /^"[^"]*"$/;
    }
    push @lines, @fields;
}
[download]

Lastly, verify your column numbers -- I passed your input line through `cut -d, -f1,2,3,4,5,9,11,12,13,15,16,17,18,19,20,21,22,23,24` and I don't seem to recall seeing the address, even after adjusting the index by removing Map,:

Axxxx,F,Bxxxx,BXXXX Axxxx,000000,NOTE:XXXXX Xxxxx (000000)@EX%SMTP:xxxxx.xxxxx@xx.com%X400:c=CA;a=;p=XX;o=xxxxxxxxxxxxx;s=XXXXX;g=Xxxxx;,Customer Srv Centre,WORKS/ SRV,SUPPORT/ SUPPORT, + 0000,(000) 000-0000 +,(000) 000-000,,,666 Snowball Paves,Vancouver,ON,CAN,J7G 3Y4

I ran another for verification:

Waaaaaa,,Lbbbbbbbb,Lccccccc Wddddddddddd,555555,SMTP:eeeeeee.fffffff@gggg.tld%HHHHHHH Iiiiiii@EX%X400:c=JJ;a=;p=KK;o=LL;s=Mmmmmm;g=Nnnnn;,Maint,,M STAT,5555,,,,,,,,,

UPDATE: Evidently the OP has a binary input file with escaped (or otherwise present) delimiters, but did not indicate which of these delimiters are escaped in the sample input.

Lastly, if it really is a binary file, don't forget to use binmode.

UPDATE: Given Corion's decision to remove any potentially personally identifiable information (a decision which I support), the above will need to be changed to reflect the new sample input... done.

During the anonymization I discovered that while coping the OP's sample input I accidentally mangled the first one -- the first `cut` output has been updated to reflect this. While the second was not mangled, it has lost context with the removal of the additional sample lines.