rosuav at gmail.com
Sat Jan 19 22:47:33 CET 2013
On Sun, Jan 20, 2013 at 4:46 AM, Tim Chase
<python.list at tim.thechases.com> wrote:
> Reading the MySQL gotchas and comparing it to the PostgreSQL gotchas, the
> MySQL ones scare the pants off my inner DBA, while the PostgreSQL ones are
> mostly "if you're running at 6+yr old version of the software, there are
> some peculiar behaviors".
Some of them are still current, though they're the least serious. The
requirement for "AS" in column aliasing is, imho, not a serious
problem - according to the SQL spec, an omitted comma is interpreted
as an alias, but PG will throw an error. The point that "SELECT
COUNT(*) FROM table" is slow is simply that MySQL happens to know the
current size of the table and can return it instantly... as soon as
you put a WHERE clause on it, both databases will perform more
comparably. Also, I think that advantage applies only to MyISAM
tables, the default but also the most problematic.
PostgreSQL does have a number of gotchas, mostly relating to
performance (for instance, the default configuration on install is
designed to work on as many systems as possible, which means it'll
perform suboptimally everywhere, and especially will not take
advantage of heaps of RAM). But MySQL has a *lot* more (whose insane
idea was it to treat "database" and "schema" as synonymous??).
More information about the Python-list