Can you give an example?

use DBI;
use strict;
use warnings;

my $csv = DBI->connect(qq'dbi:CSV:f_dir=.;csv_eol=\n') or die;

my $d = $csv->selectall_arrayref(qq(select * from longread where a = ?
+), {}, 1);

print length $d->[0][1], "\n";
[download]

absolutely: here is the first 10 lines of text: http://pastebin.com/R8E6jVhs

Here is the code:

#! c:\program files\perl\bin
use DBI;
my $csv="viewmanifest.csv";
$dbh=DBI->connect('DBI:CSV:$csv') or die "Can't Connect";
$sth=$dbh->prepare("SELECT RowId, dahandle, darelativepath, dafilestat
+e, dafilestatetime, dafilesource, fileguid, size, filemd5, child, chi
+ldtype, childhandle, container, containerfolder, maghandle, mailconta
+inererror, mailcontainer, mailfolder, mailfolderpath, contentmd5, cre
+atedtime, datescanned, embeddedchildren, docclass, handle, docnum, os
+folder, filename, docext, origdocext, filemode, fullpath, filetype, a
+uxfiletype, lastaccesstime, lastmodifiedtime, lastchangetime, mimetyp
+e, kftdesc, ocrpath, ocrstatus, ownergroup, owneruser, acls, parent, 
+parsingstatus, parserversion, altbody, importpath, importdevice, impo
+rtarea, ancestordahandle, depotfile, inode, batch FROM $csv") or die 
+"Can't prepare";
$sth->execute();
$sth->dump_results();
$sth->finish();
$dbh->disconnect();
[download]

Here are the results for the first line (notice the ellipses)

'0', '000083e0143381bbe8db4b0db7baa5...', 'vmi/v920m3/weekly.20091014/pro...', 'imported', '2011-04-15-21-00-51', '000083e0143381bbe8db4b0db7baa5...', '', '0000000000000063906', 'a088be10e6f02f614385fa19628cfb6f', '', '', '', '', '', '', '', '', '', '', '', '', '2011-04-15-21-00-51', '', 'loose', 'a088be10e6f02f614385fa19628cfb...', '', 'vmi/v920m3/weekly.20091014/pro...', 'en_X11.xml', 'xml', '', '33188', '', 'extensible markup language (xml)', '', '2009-10-15-19-15-34', '2009-10-15-19-15-49', '2009-10-15-19-15-49', '', '', '', '', '100', '13936', '', '', '00000 SUCCESS', '3.4;3.4', '', '000083e0143381bbe8db4b0db7baa5...', '', '', '', '', '32457187', 'f8023f8fc92a4fc4'

Other things of note, could have unicode filenames...

dump_results is not meant to fetch data!

use strict;
use warnings;
use autodie;

use DBI;
use Data::Peek; # For debugging only

my $dbh = DBI->connect ("dbi:CSV:", undef, undef, {
    f_dir      => ".",
    f_ext      => ".csv/r",
    RaiseError => 1,
    PrintError => 1,
    });

my $sth = $dbh->prepare ("select * from viewmanifest");
my %rec;
$sth->execute;
$sth->bind_columns (\@rec{@{$sth->{NAME_lc}}});
while ($sth->fetch) {
    DDumper \%rec; # For debugging only
    print $rec{dahandle}, "\n";
    }
$dbh->disconnect;
[download]

FWIW DBD::CSV doesn't need to support LongReadLen as the only limit to the size of the fields is implied by the available memory and the remaining space on disk (DBD::CSV is always AutoCommit so the data in the table should also fit on disk). The other database that I know of that ignores LongReadLen completely as it has direct bindings to the perl variable is DBD::Unify.


Enjoy, Have FUN! H.Merijn