[R-sig-DB] Parameterised queries

[Paul Gilbert]

(Possibly naive question.)

>The reason inserts are slow...

I've now seen this stated a few times on this thread and I don't fully 
understand the context everyone seems to be assuming. With a Perl client 
I can insert multiple records with a single insert command, over a 
network, and it seems to be nearly as fast as loading with the batch 
utility locally on the server. As I recall, I had to block less that 
10,000 records or something choked, but multiple blocks loaded a fairly 
large database reasonably quickly. (I'm travelling and cannot verify 
this number at the moment.) It is import to load before specifying 
indexes, otherwise the index gets recalculated with each insert.

When people say inserts  are slow:

-Is this an R specific problem?
-Is it assumed indexes are already specified?
-Or am I missing something else once again?

Thanks,
Paul