You can make the timestamps unique by appending a sequence number. A 32 bit Unix timestamp can map into a 64 bit Sqlite integer.
P Kishor wrote: > On 4/14/08, Chris Tracy <[EMAIL PROTECTED]> wrote: >> Err, >> >> Make that example table data: >> >> <...> >> >> 1167611400|89|7|34 >> 1167611400|101|5|4 >> 1167611400|147|9|14 >> 1167611400|173|7|2 >> 1167611400|195|8|1 >> 1167611400|314|8|12 >> 1167611400|347|6|48 >> 1167611400|364|1|49 >> >> <snip> >> >> 1167615600|111|7|39 >> 1167615600|155|2|8 >> 1167615600|157|4|7 >> 1167615600|220|4|47 >> 1167615600|247|7|21 >> 1167615600|261|8|30 >> 1167615600|308|9|20 >> >> <...> >> >> As I'd originally said, the timestamps are inserted in ascending >> order. (I just screwed up pasting my example values in the wrong order) >> >> > > > ok, but first, I don't really understand the following -- > >> SELECT * from bridge_table WHERE timestamp = <someval>; >> >> and almost immediately get back all the results I want, since they're >> right next to each other in the db. However, the actual SELECT doesn't >> return for almost 13 additional seconds, as the entire table is scanned >> for other rows where timestamp might be <someval> > > Your first sentence above is that you "almost immediately get back all > the results you want" then you go on to say that the "actual SELECT > doesn't return for almost 13 additional seconds." So, what is it > exactly that you "almost immediately get back"? > _______________________________________________ > sqlite-users mailing list > [email protected] > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
