The most important thing I know about performance is that until you
know with proof from tests you're just guessing, and even with lots of
experience acting on a guess can hurt as much as help. In fact, ask
anyone with a lot of profiling experience and they'll know based on
lots of wrong guesses that it is useless to try to optimize without
tests, and also without tests you never know if you did any good or
not (or did any harm...).
So, if this is something that interests you I would invite you to do
some testing and report the differences in performance and storage
impact for a few different scenarios. Based on that we can have a real
discussion about it and make some decisions.
-David
On Feb 9, 2009, at 2:22 AM, madppiper wrote:
Hey everyone,
while taking a closer look at the fieldtypemysql.xml file i noticed
that we
are using predominantly varchars, which got me thinking. I know out
of my
own experience (and from several authors), that the use of dynamic
datatypes
results in increased query times. A well formatted database table
may in
times be up to 6x the original speed on update & delete functions...
From my point of view, there is really no need for us to stick to
varchars
and blobs anyway. We are already delimiting the size of the varchar
field
(20 chars, 100m or 255) and could do the very same with chars as
well. I
know that this may result in bigger overall filesizes for the
database and I
am fully aware of that downfall, but I think that we could gain alot
from
this change...
Also: I know that this only takes effect if we change ALL the
dynamic fields
to static ones, a combination of both will not really have an effect
at all.
So what are your thoughts on this?
Cheers,
Paul
--
View this message in context:
http://www.nabble.com/Discussion%3A-Mysql-Db-Performance-tp21909393p21909393.html
Sent from the OFBiz - Dev mailing list archive at Nabble.com.
