Re: Re: Speeding up the DBI

[tantarbobus]

in reply to Re: Speeding up the DBI
in thread Speeding up the DBI

Yes it is. DBD::mysql behaves in the same way that DBD::Pg does when it comes to pulling all of the rows from the database server.

If you want DBD::mysql to act as if it were using cursors, you can specify mysql_use_rows on the statement handle, and then it will not pull over the whole data set on execute.

For DBD::Pg, you can declare a cursor and then execute the fetch in a loop until you have recieved all of the data. Admittedly this coould be done in a cleaner fashion where prepare could take an attribute and if prepare sees that attribute it will declare a cursor and rewrite your statement to use the cursor (which is what Pg::Simple does,IIRC)

[diotalevi]

I had been under the impression that cursors don't work through DBD::Pg. I'm mistaken?

[tantarbobus]

$dbh->do(q{DECLARE imprecator CURSOR FOR SELECT * FROM tablename}) ;
my $number = 10;
my $fetch = $dbh->prepare(q{FETCH $number FROM imprecator});


##
### Fetch first $number * 10 rows.
##
while ($fetch->execute()) {
     Data::Dumper::Dumper($fetch->fetchall_arrayref());
     last if $fetch->rows() < $number;
}

Note: The code above has not been tested; it is for illustrative purposes only.