On 27.09.2011 14:47, Jeroen Vermeulen wrote: > On 2011-09-27 19:26, Abel Deuring wrote: > >> I wrote a bit about StormRangeFactory in >> https://dev.launchpad.net/Database/Performance > > Thanks for that. I made some very small updates, on the principle that > I should leave what I use slightly better than I found it. > > One thing I'm not sure about: the text said that _calculating_ offsets > is expensive. That's probably not exactly what was meant ("next_offset > = offset + batch_size" can't be too expensive :), so I changed it to > "finding the end of the result set, and skipping to the right offset."
Right, thanks for the "improved precision" :) > > Those would correspond to the counting query for the batch navigation > widget and the OFFSET argument in the main query. But are they both > part of the problem that StormRangeFactory solves? yes, this is the core of what StormRangeFactory is intended to improve: Get rid of the need to count larger result sets for OFFSET and SELECT count(*). (The latter is replaced by an estimation. It might make sense to change the text "15->20 of 234 results" with something like "15->20 of approximately 234 results" -- I suspect that the estimated result size can be drastically wrong for non-trivial queries.) _______________________________________________ Mailing list: https://launchpad.net/~launchpad-dev Post to : launchpad-dev@lists.launchpad.net Unsubscribe : https://launchpad.net/~launchpad-dev More help : https://help.launchpad.net/ListHelp
