RE: update? insert? Perl script

[Michael Widenius]

>>>>> "Jay" == Jay Miller <jaymiller@stripped> writes:

Jay> In a case like this we are using REPLACE, which performs a DELETE and
Jay> INSERT.  This is easier than checking first and either UPDATEing or
Jay> INSERTing.  I'm sure it places a higher load on the MySQL server, but it
Jay> takes some off of the client and possibly reduces some network traffic.

Jay> Any thoughts?

Hi!

The load of REPLACE should be much less than an SELECT / DELETE + an
INSERT.

REPLACE is implemented as an INSERT which is rolled back in case of
duplicated keys. The duplicated row is then deleted and then we try
another INSERT..

Regards,
Monty