Re: [sqlite] Best approach for xBestIndex/xFilter effectiveness

2011-03-22 Thread Dan Kennedy
On 03/23/2011 01:07 AM, Jay A. Kreibich wrote: > On Tue, Mar 22, 2011 at 06:25:04PM +0700, Dan Kennedy scratched on the wall: > >> SQLite assumes that the result of each expression in the WHERE >> clause depends only on its inputs. If the input arguments are >> the same, the output should be do.

Re: [sqlite] Best approach for xBestIndex/xFilter effectiveness

2011-03-22 Thread Jay A. Kreibich
On Tue, Mar 22, 2011 at 06:25:04PM +0700, Dan Kennedy scratched on the wall: > SQLite assumes that the result of each expression in the WHERE > clause depends only on its inputs. If the input arguments are > the same, the output should be do. Since random() has no inputs, > SQLite figures that it

Re: [sqlite] Best approach for xBestIndex/xFilter effectiveness

2011-03-22 Thread Max Vlasov
On Tue, Mar 22, 2011 at 2:25 PM, Dan Kennedy wrote: > On 03/22/2011 04:26 PM, Max Vlasov wrote: > > Hi, > > > > recently I finally started experimenting with virtual tables and there's > at > > least one thing I can not understand. > > > > As I see xBestIndex/xFilter were

Re: [sqlite] Best approach for xBestIndex/xFilter effectiveness

2011-03-22 Thread Dan Kennedy
On 03/22/2011 04:26 PM, Max Vlasov wrote: > Hi, > > recently I finally started experimenting with virtual tables and there's at > least one thing I can not understand. > > As I see xBestIndex/xFilter were developed to allow fast searching if the > implementation is able to do this. But there's