Hello, ok here's as short and succinct a sample as I can create.

use Text::CSV;
use CGI;

my($row) = "search/¿Cuales son las partes de una cadena de conexión??s
+cope|ids_jdbc_011.htm|0|1|1|0";

my $csv = Text::CSV->new ({ binary => 1, sep_char => "|" });
my $q = new CGI;

# print the HTML header and start html
print $q->header;
print $q->start_html;

 # first, print $row as is
 print $q->p("ROW: $row");

 # next, parse with $csv
 $csv->parse($row);
 my @els = $csv->fields;

 # print the first field
 # this displays the black diamond ? symbol for ¿ and ó
 print $q->p("CSV Parse, field 0:",$els[0]);
 
 # split instead
 my(@splits) = split('\|',$row);
 
 # print the first element in @splits.
 # As noted, this one displays properly in the browser.
 print $q->p("split 0:", $splits[0]);
 
print $q->end_html;
 
exit;
[download]

thanks

======================

UMM, update, when I actually ran the above in my http server I got the opposite results, but with weird errors.

ROW: search/¿Cuales son las partes de una cadena de conexión??scope|
+ids_jdbc_011.htm|0|1|1|0

CSV Parse, field 0: search/¿Cuales son las partes de una cadena de con
+exión??scope

split 0: search/¿Cuales son las partes de una cadena de conexión??sc
+ope
[download]

Paint me confused.

In the real script, $row is coming from a @sorted_array from an SQL query. This is getting confusing so maybe I should withdraw my question.