Jeff,
I'm adding more support for csv in DBI::Shell. However, I'm
seeing some odd results.
If I use this connection string and query. 0 Rows are fetched.
(some code removed for brevity)
my $dbh = DBI->connect("DBI:CSV:f_dir=.")
...
my $sth = $dbh->prepare("select * from myfile")
...
$sth->execute() ...
my $count = 0;
while (my $row = $sth->fetchrow_arrayref) {
print join "|", @$row, "\n";
$count++;
}
warn "Fetched $count rows\n";
Where as this return rows 3 of 4 rows (first row is consumed for column
names?)
my $dbh = DBI->connect("DBI:CSV:f_dir=.;csv_eol=\n")
...
The default behavior (from the docs) looks carriage return, line feed.
csv_csv The attributes csv_eol ... Defaults are "\015\012"
Would it be more portable if the default eol be based on the OS eol?
Tom
Perl : 5.008 (i386-linux-thread-multi)
OS : linux (2.4.20-2.48smp)
DBI : 1.38
DBD::mysql : 2.1021
DBD::Sponge : 11.09
DBD::Proxy : 0.2004
DBD::Pg : 1.31_3
DBD::Oracle : 1.14
DBD::Multiplex : 0.9
DBD::File : 0.2001
DBD::ExampleP : 11.10
DBD::CSV : 0.2002
SQL::Statement : 1.005
File: out.pl
#!/usr/bin/perl
use strict;
use warnings;
use DBI;
# Returns 0 rows.
# my $dbh = DBI->connect("DBI:CSV:f_dir=.")
# Works, except skips the first line (fields names)
my $dbh = DBI->connect("DBI:CSV:f_dir=.;csv_eol=\n")
or die "Cannot connect: " . $DBI::errstr;
$dbh->{RaiseError} = 1; $dbh->{PrintError} = 1;
my $sth = $dbh->prepare("select * from myfile")
or die "Cannot prepare: " . $dbh->errstr();
$sth->execute() or die "Cannot execute: " . $sth->errstr();
my $count = 0;
while (my $row = $sth->fetchrow_arrayref) {
print join "|", @$row, "\n";
$count++;
}
warn "Fetched $count rows\n";
$sth->finish();
$dbh->disconnect();
exit;
__END__
Contents of myfile
1,Tom
2,Aaron
3,Ryan
4,Susan
