Hi Niels, Thanks for your answers. On Saturday 15 September 2007 14:53, Niels Nes wrote:
It seems your benchmark is mostly measuring the communication overhead of both database systems. Indeed thats where MySql excels.
Are you saying the MySQL communication overhead is generally lower (eg more efficient protocol and/or client libs) or just that their PHP extension is better optimized (for this use case)? Do you know if MonetDB supports unix domain sockets, as postgres and mysql do? On a local machine, that should prover faster.
Add some more complexitity to the queries (joins and/or aggregates) and MonetDB should behave just fine. Maybe even changing your query from
"select col1 from big_table where col1 = $i";
into
"select count(*) from big_table where col1 = $i";
I'll try both an aggregate and a join, separately. thanks for the suggestion.
may allready change the picture. For sure your big_table isn't that big.
Do you mean "not big" in terms of number of rows, or columns, or both? With these very simple select queries, my expectation would be that I would see the same linear trend if the number of rows were 10000 or 10,000,000. regards, -- Dan Libby Open Source Consulting San Jose, Costa Rica http://osc.co.cr phone: 011 506 223 7382 Fax: 011 506 223 7359