Performance Question

Hello List:

I've been told that an update to a record is equivalent to a delete and insert
operation. We have a utility written in Perl that brings into sync certain
elements of 50 thousand records on 8 structurally identical databases. We
threw together the script and decided to just delete the record and re-insert
it with the data that was brought into sync. Now the question: Is it just as
fast to do it this way, or is there some hidden advantage to performing an
update?

Just curious.

TIA

master=# select version();
version
--------------------------------------------------------------------------------------------------------------
PostgreSQL 7.4.6 on i686-redhat-linux-gnu, compiled by GCC gcc (GCC) 3.2.3
20030502 (Red Hat Linux 3.2.3-49)