Re^2: DBD::CSV::prepare() problems

[polelo]

in reply to Re: DBD::CSV::prepare() problems
in thread DBD::CSV::prepare() problems

Dear Tux

Thanks for your reply

this is how I connect to the CSV database

use DBI;
...

#MS Windows path to files
my $fpath="G:\\in";
#LINUX server paths
#my $fpath="/users/excel/in";

my $csvfile="Trx_24Aug2012_0409.csv";
my $cellh = DBI->connect("DBI:CSV:f_dir=$fpath;csv_eol=\n;");
    
#associate csv file with tablename
my $fname=$fpath.'/'.$csvfile;
$cellh->{'csv_tables'}->{'CELLS'} = {'file' =>$fname};
    
my $sql="select * from CELLS";
my $stcellh =  $cellh->prepare($sql) or die "$DBI::errstr\n";
      
$stcellh->execute() or die "$DBI::errstr\n";
my $colunmcount = $stcellh->{NUM_OF_FIELDS};
print "$colunmcount\n\n";
$stcellh->finish();

Thanks again

Polelo

[remiah]

Hello Polelo


I tried with my FreeBSD but it says nothing for filename. No error message, no warnings.

#!/usr/bin/perl
use strict;
use warnings;
use DBI;

my $dsn="dbi:CSV:";

my $dbh=DBI->connect($dsn, "","",{
        f_encoding       => "utf8",
        f_dir            => "./",
})or die DBI->errstr;

$dbh->{csv_tables}->{t1} = {
        'file' => '17_Trx_22Aug2012_0405.csv',
        'eol' => "\n",
};

my $sth;
$sth=$dbh->prepare("select * from t1");
$sth->execute;
while( my $r = $sth->fetchrow_hashref){
        print "$r->{item1},$r->{item2}\n";
}
$sth->finish;
$dbh->disconnect;

#print version
print "perl version =$]\n";
for (qw/DBI Text::CSV_XS SQL::Statement DBD::CSV/){
        print "$_ ==>".$_->VERSION."\n";
}

versions of mine. How is yours?


perl version =5.012002
DBI ==>1.615
Text::CSV_XS ==>0.85
SQL::Statement ==>1.33
DBD::CSV ==>0.35

[Tux]

Because you use the file => $filename approach, you do not use the case-insensitiveness from SQL::Statement.

I do not think you should define csv_eol

When on Windows *and* other OS's, do not use $dir."/".$file, but use File::Spec.

When in doubt about casing, use something like

chdir $f_path;
foreach my $file_name (glob "*.*") {
    lc $file_name eq lc $csvfile or next;
    $csv_file = $file_name;
    last;
    }

my $dbh = DBI->connect ("dbi:CSV:", undef, undef, {
    f_dir      => $f_path,
    f_ext      => ".csv/r",
    f_encoding => "utf-8",
    f_schema   => undef,
    RaiseError => 1,
    PrintError => 1,
    });
$cellh->{csv_tables}{cells} = { file => $csv_file };

my $sth = $sbh->prepare ("select * from cells");
$sth->execute;
print "fields: @{[@{$sth->{NAME_lc}}]}\n";

untested, but should work.

[Anonymous Monk]

When on Windows *and* other OS's, do not use $dir."/".$file, but use File::Spec.

Go ahead and use / on every OS you have

[afoken]

Go ahead and use / on every OS you have

Nonsense! Not every OS uses the forward slash (/) as path separator.

DOS and Windows (and OS/2, I think) allow to use both backslash (\) and forward slash, so you can use the forward slash on Windows, except when you try to use external tools that expect that you only use the backslash.

On a classic, non-X MacOS, you will fail miserably because MacOS uses single colons (:).

Same game, different rules with VMS: / works only in Unix emulation, the native interface uses single colons, double colons (C<::>), square brackets ([]), angle brackets (<>), periods (C<.>), and semicolons (C<;>) to separate the parts of a fully qualified file name.

File::Spec is not perfect, but it still is better than simply using the forward slash everywhere.

Alexander