Hi, Le 11/07/2013 11:20, Martin Kersten a écrit :
Without insight into your benchmark queries and database schema for the row stores, little can be said.
Sure... but I just wanted to get past the obvious problems before going into the details.
A typical pitfall could be to use row-based queries such as "SELECT * FROM tableexpr", which would call for expensive tuple reconstruction of all columns. No, we don't. In fact, Mondrian doesn't, to be precise. Another explanation can come from a multidimensional index maintained in a rowstore to speedup grouping. We did not declare any index in MonetDB. Maybe this is the problem ? Last but not least, it could be grouping order or skewed data distribution.
MonetDB provides some tools to see where time goes, which may give a hint. http://www.monetdb.org/Documentation/Manuals/MonetDB/Profiler/Stethoscope I will investigate and try to find out what is happening. I'll come back with more information.
Your table sizes reported is not extreme at all. That is what I thought... and why I'm asking, as my current results are really not expected...
Regards, Franck