Unterminated `s' command

Question

I am working on a perl script that finds will take customer service information and change the service ids. The code i have now can take a 2 column csv file for reference, and swap the numbers, but when i try to use "\n" or "\|" like i need to, it will give me the `ol:

"sed: -e expression #1, char 29: unterminated `s' command"

Here's my original code that works for changing JUST the number within the quotes:

#!/usr/bin/perl
use strict;
use warnings;
use Text::CSV; #load Text::CSV module for parsing CSV
use LWP::Simple; #load LWP module for doing HTTP get

my $file = 'sid_swap.csv'; #CSV file to parse and load
my $csv = Text::CSV->new(); #create a new Text::CSV object

open (CSV, "<", $file) or die $!; #open CSV file for parsing

while (<CSV>) {
    if ($csv->parse($_)) {

        my @columns = $csv->fields(); #parse csv files and load into an array for each row
        my $newSID = $columns[0];
        my $oldSID = $columns[1];
        system("sed -i 's/\<row SERVICE_MENU_ID=\"$oldSID\"/\<row SERVICE_MENU_ID=\"$newSID\"/g' customer_data.txt");

    } else {
        my $err = $csv->error_input;
        print "Failed to parse line: $err";
    }
}
close CSV;

And here is the new code that throws the error:

#!/usr/bin/perl
use strict;
use warnings;
use Text::CSV; #load Text::CSV module for parsing CSV
use LWP::Simple; #load LWP module for doing HTTP get

my $file = 'sid_swap.csv'; #CSV file to parse and load
my $csv = Text::CSV->new(); #create a new Text::CSV object

open (CSV, "<", $file) or die $!; #open CSV file for parsing

while (<CSV>) {
    if ($csv->parse($_)) {

        my @columns = $csv->fields(); #parse csv files and load into an array for each row
        my $newSID = $columns[0];
        my $oldSID = $columns[1];
        system("sed -i 's/\<row SERVICE_MENU_ID=\"$oldSID\"/\n$newSID\|/g' customer_data.txt");

    } else {
        my $err = $csv->error_input;
        print "Failed to parse line: $err";
    }
}
close CSV;

Thanks for any help you can provide!

Answer 1

To debug, change system("sed ...") to die("sed ...");. You'll see you're what trying to execute

sed -i 's/<row SERVICE_MENU_ID="OLDSID"/
NEWSID|/g' customer_data.t

I guess sed doesn't like actual newlines in the middle of its arguments? That can be fixed using proper escaping, but your approach is... insane. You're processing the entire file for each row of the CSV!

open(my $CSV, "<", $file) or die $!;    # Let's not use global vars!

my %fixes;
while (<$CSV>) {
    $csv->parse($_)
        or die "Failed to parse line: ".$csv->error_input."\n";

    my ($newSID, $oldSID) = $csv->fields();
    $fixes{$oldSID} = $newSID;
}

my $pat = join "|", map quotemeta, keys(%fixes);
my $re = qr/$pat/;

{
    local @ARGV = 'customer_data.txt';
    local $^I = '';  # "perl -i" mode
    while (<>) {
        s/<row SERVICE_MENU_ID="($re)"/\n$fixes{$1}|/g;
        print;
    }
}

Doing it in one pass also solve the problem of

1111,2222
2222,3333

being the same as

1111,3333
2222,3333