On Tue, Oct 2, 2012 at 5:12 PM, Duquette, William H (318K) < william.h.duque...@jpl.nasa.gov> wrote:
> Howdy! > > The SQLite3 Tcl interface has a "nullvalue" command, which determines how > NULLs are represented as Tcl values. If you do a query on a NULL value, > you get the "nullvalue" value. ("nullvalue" defaults to the empty string.) > > However, if a NULL value is passed to a custom SQL function, defined using > the Tcl [$db function] command, the function body (a Tcl proc) appears to > get the empty string instead of the desired "nullvalue" for the given > value. This seems like a bug to me. Am I wrong? > Please use this fix: http://www.sqlite.org/src/info/9bf64b661 The same patch is also on trunk (http://www.sqlite.org/src/info/c1f10a2643). However, I'm in the middle of a significant query planner enhancements on trunk - specifically I'm working to make the query planner much more aggressive about optimizing out ORDER BY clauses and arranging for the query results to naturally come out in the desired order instead of having to sort. You are, of course, welcomed to try trunk instead. Trunk currently passes all regression tests. And it would actually be useful for us if you tried out trunk and let us know about any ORDER BY anomalies that you encounter. Just be aware that the query planner on trunk is under active development and might not be quite as stable as the 3.7.14 release. > I'm using SQLite3 3.7.7.1; the function is called from a trigger (if that > matters). > > Will > > -- > Will Duquette -- william.h.duque...@jpl.nasa.gov > Athena Development Lead -- Jet Propulsion Laboratory > "It's amazing what you can do with the right tools." > > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > -- D. Richard Hipp d...@sqlite.org _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users