RE: : RE: bytes vs chars

No need for further comment on the length discussion - as with many general principles, people who think about what and why tend to do it right, people who don't pause for thought tend to do it wrong. Being part of the discussion virtually guarantees that one falls into the former group.


As far as testing goes, though, I would like to agree with you that this is the sort of problem that should be detected - even before the test to scale, of course - but (a) my experience has been that few sites actually know how to design proper tests and (b) it's often the case that a test displays a warning symptom but the "fix" is aimed in completely the wrong direction - for example "it's generating a lot of redo, let's put the online logs onto SSD", rather then "it's generating a lot of redo - surely it's generating far more redo than it should be, why that happen".



Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
_at_jloracle

________________________________________
From: oracle-l-bounce_at_freelists.org [oracle-l-bounce@freelists.org] on behalf of Dave Morgan [oracle@1001111.com]
Sent: 17 March 2016 14:49
To: oracle_at_1001111.com; Oracle-l
Subject: Re: : RE: bytes vs chars



However, I still go to my claim that this is the sort of problem that should be detected in testing. If I am pounding in 1 or 10 rows per second then the excessive
resource usage should be obvious and must be fixed. If I am inserting 1 row per hour should I worry about it?

--