>I'm not even sure it makes a difference. The MYSQL DBI driver lets you run
>multiple distinct queries at the same time, and as I understand the MYSQL
>client library, you can't really do that without having multiple
>connections. So I think that MYSQL's DBI driver is pulling the entire
>result set over on the fetch.
>Of course, I could be wrong, I haven't dug deeply to figure it out
>specifically. I'd love if someone were willing to post a more explicit
The MySQL DBD defaults to using the mysql_store_result() call from the C API
on which it's built. There are two DBD-specific attributes you can manipulate
to control this:
if you set this, mysql_store_result() is used to fetch the result set
(all rows at once).
if you set this, mysql_use_result() is used to fetch the result set
(rows not fetched until you call for them).
You get the usual advantages and disadvantages from setting these attributes
that you do when using the C API calls directly in a C program.
Paul DuBois, paul@stripped