On 7/11/06, Jan Wieck <JanWieck@xxxxxxxxx> wrote:
On 6/30/2006 1:07 PM, Merlin Moncure wrote: I have no clue what flushing tables with lock might be good for. Are applications in MySQuirreL land usually smarter than the DB engine with respect to when to checkpoint or not?
no, but the ability to flip a database into read only mode with such an easy command is relatively useful. i very much pg had a read only mode, btw.
The multiple insert stuff is not only non-standard, it also encourages the bad practice of using literal values directly in the SQL string versus prepared statements with place holders. It is bad practice because it introduces SQL injection risks since the responsibility of literal value escaping is with the application instead of the driver.
good points, and pg can accomplish similar via insert select union all, etc., but mysql version of same is better syntax imo, so i guess i should take it up with the sql standard. As to preparing statements, I agree in principle although I don't know if that is a good argument not to make the non-paramaterized interface more powerful.
Everything that teaches new developers bad things counts as a disadvantage in my book, so -1 on that for MySQL too.
no comment. :) small disclaimer: I am right now administrating a relatively large mysql database infrastructure which I inherited. I was hired with for the express purpose of converting it to pg. meanwhile since writing the op I got burned really badly by the mysql replication where the slave became off synch with master on an important table, something you might appreciate. that said, i tried to put fairness in my comparison, many pg/mysql comparisons ulimately resort to a dismissive mysql diss which does not play well to the uninformed third party. so, I made an attempt at something with some substance. regards, merlin
