On 09/27/2012 10:37 PM, Chris Angelico wrote:>[...] > * MySQL is designed for dynamic web sites, with lots of reading and > not too much writing. Its row and table locking system is pretty > rudimentary, and it's quite easy for performance to suffer really > badly if you don't think about it. But if your needs are simple, MySQL > is probably enough. PostgreSQL uses MVCC to avoid locks in many cases. > You can happily read from a row while it's being updated; you'll be > unaware of the update until it's committed.
MVCC comes with a cost though, as anyone who has ever needed to do a SELECT COUNT(*) on a large Postgresql table knows. >[...] > * Both engines have good support in popular languages, including > (dragging this back on topic, kicking and screaming) Python. Maybe things are different now but a few years ago I was trying to choose between Postgresql and Mysql about the time Python 2.4 (I think) was released. After waiting for over a year for the Python mysql dbi module to be released for the then current version of Python (I needed a binary for Windows) I finally gave up and decided to go with Postgresql (the psycopg2 module was available a very short time after the new Python was.) -- http://mail.python.org/mailman/listinfo/python-list
