On Sep 19, 2012 2:32 AM, "Fabiano" <fabiano@...> wrote:

>
> Hi all!

Hi!

>
> Today it seems easier and >simpler for me to create all >character fields

encoded in UTF8.

IMHO it's not easier :-)

>
> Is there something against doing >this?

At the cost of sounding trivial, the main problem I see with that lies with
fields that explicitly need to be ASCII: very few these days, yet there
nonetheless.

>
> I asked some days ago about >index sizes in this post (http://>

tech.groups.yahoo.com/group/>firebird-support/>message/119234),

>and it seems UTF8 can generate >indices keys up to 4X larger than >a usual

key.

Not sure about that, however I struggle to see how that can be a problem
with today's RAM sizes.

>
> As the majority of characters stored in these fields are below code 127

in ASCII table, with some rare accents and special symbols, i would like to
know if using UTF8 should increase too much my database or affect its
performance.

It will certainly have an effect on size and probably performance. I don't
know - however - if the performance one will be measurable, meaning that it
seems to me you might really notice a degradation if you are doing a select
* of millions of records without a where which would be wrong anyway.
I don't think the degradation will be noticeable otherwise. There are - of
course - edge cases but they should not modify the general understanding of
this.

> Regards,

Rgds,

> Fabiano

A
(argh deleting all of that stuff using the phone backspace is AWFUL...
scarred me for life heaven's sake!)


[Non-text portions of this message have been removed]