Comment Re:Performance and replication? (Score 1) 213
> Performance. Specifically, how does it compare with Mysql? Some real app comparison would be very nice.
I made some smaller tests about this. As general rules:
* for complex queries, Postgres is faster.
* for lot's of parallel reads/writes, Postgres is faster.
* for big databases, PostgreSQL is faster.
* for single and non-concurrent writes, MySQL is faster.
* for simple queries and small databases, MySQL is a little bit faster.
For single writes I don't need top speed. So I don't need MySQL ... ;-)
One example is a Select with GROUP BY on a VARCHAR column on a table with >500000 rows. MySQL takes > 3 seconds on a MyISAM Table, locking the whole table for write access. Postgres does the same *without* lock (write access is possible) in about 2 seconds. MySQL with innodb (no locking) takes more then 10 seconds. All on the same data, same machine and comparable memory configuration.
MySQL and MyISAM is the hell when you have some expensive queries and want to write permanently to the same table or in general: when you have complex queries.
Ciao
Alvar, http://www.a-blast.de/statistik/ (still with MySQL :-( )
I made some smaller tests about this. As general rules:
* for complex queries, Postgres is faster.
* for lot's of parallel reads/writes, Postgres is faster.
* for big databases, PostgreSQL is faster.
* for single and non-concurrent writes, MySQL is faster.
* for simple queries and small databases, MySQL is a little bit faster.
For single writes I don't need top speed. So I don't need MySQL
One example is a Select with GROUP BY on a VARCHAR column on a table with >500000 rows. MySQL takes > 3 seconds on a MyISAM Table, locking the whole table for write access. Postgres does the same *without* lock (write access is possible) in about 2 seconds. MySQL with innodb (no locking) takes more then 10 seconds. All on the same data, same machine and comparable memory configuration.
MySQL and MyISAM is the hell when you have some expensive queries and want to write permanently to the same table or in general: when you have complex queries.
Ciao
Alvar, http://www.a-blast.de/statistik/ (still with MySQL
