> On 28 Jul 2014, at 10:58pm, Joe Mistachkin <sql...@mistachkin.com> wrote: > > Eric DAVID wrote: > >> Yes, it is this one. With the new script I gave in my last mail it has >> changed to >> >> SELECT * FROM Pylones P >> JOIN Lignes L ON L.lig_id = P.lig_id >> JOIN idx_Pylones_pyl_geom I ON P.pyl_id = I.pkid WHERE L.lig_id = 1; > > So far, I'm unable to reproduce your results here. > > Could you please provide the list of all SQL statements used to get the > database into the state where the issue is seen, including any INSERT > statements used to populate it with the test data.
I'm beginning to wonder whether there's a problem with the handling of the floating point numbers used for this data. A number with ten significant digits like these: 917130.8125 917147.4375 could run up against a slop or fiddle factor when assessing equality or inequality. Solving the problem could involve testing in the same programming language that the original user used. Original Poster: if you take the list of commands you showed us in your earlier post and enter them straight into the SQLite shell tool (to create a new database) do you get the same results as you get when you execute them in your own program ? Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
