More efficient, which?

[Zurev]

Just wondering, the easiest way I can think of to give an example is this:

Scenario A:
Need to figure out post count on a forum
Everytime someone posts their row updates a post_count field by 1, and then that number gets displayed.

Scenario B:
You use count();(If I remember correctly) to tally all the posts that are started by that user and display that number.

Since most major forum systems use a post_count field in their tables I'm going to assume scenario A is more efficient, but does it really matter?

[tripy]

but does it really matter?

On really large tables, it can.

Mysql with myIsam tables use the counter trick to lighten the server work.
I encountered situations where having multi million rows to be counted, with a where clause led to several seconds of work for the db (postgresql, back in 1999. They got better since..)

So yes, it can make a difference.

[r937]

how many times does a person post, versus how many times is his post count retrieved for display?

that'll guide your decision