Usually not
Well, maybe so but you really should limit in the case of postgres.
This is basically your example code running against 9.4devel, with and without a limiting where-clause:
$ perl ./tux.pl #
PostgreSQL 9.4devel_HEAD_20140502_2044_0717748 on x86_64-unknown-linux
+-gnu, compiled by gcc (GCC) 4.9.0, 64-bit
prepare execute finish
no where 0.00007 5.52991 0.02707
no where 0.00005 5.42172 0.03283
no where 0.00005 5.42320 0.03247
where 0=1 0.00005 0.00049 0.00000
where 0=1 0.00002 0.00013 0.00000
where 0=1 0.00002 0.00012 0.00000
(foo has 10M 1-column rows; just a
create table foo as select n from generate_series(1, 10000000) as f(n);
)
(What the hell -- let me just dump the test here too, even if it's a bit clunky (disks are cheap and patient):
use strict;
use warnings;
use DBI;
use Time::HiRes qw/gettimeofday tv_interval/;
my $dbh = DBI->connect or die "oops - $!\n";
print $dbh->selectrow_arrayref('select version()')->[0], "\n\n";
my $sql1 = "select * from foo";
my $sql2 = "select * from foo where 0 = 1";
print " prepare execute finish\n";
time_this( $dbh, $sql1, ' no where' );
time_this( $dbh, $sql1, ' no where' );
time_this( $dbh, $sql1, ' no where' );
print "\n";
time_this( $dbh, $sql2, 'where 0=1');
time_this( $dbh, $sql2, 'where 0=1' );
time_this( $dbh, $sql2, 'where 0=1' );
sub time_this {
my ($dbh, $sql, $how) = @_;
my $t0;
$t0 = [gettimeofday];
my $sth = $dbh->prepare( $sql );
print $how, " ";
printf(" %7.5f", tv_interval($t0 , [gettimeofday]));
$t0 = [gettimeofday];
$sth->execute;
printf(" %7.5f", tv_interval($t0 , [gettimeofday]));
my @fld = @{$sth->{NAME}};
$t0 = [gettimeofday];
$sth->finish;
printf(" %7.5f", tv_interval($t0 , [gettimeofday]));
print "\n";
}
) | 