Igor: As usual, you're right.
The only problem is that I think that generally, the type of SQL that you use is not too common, at least among those of us who do not have much practice with the language. At least I tend to get that type of expressions away of the SQL query. And I suppose that in my case, the C++ compiler is faster that the SQL engine. Any way, again from a theoretical point of view you are right. If you could inspect my code I'm sure that you become horrified with the convolutions I need because my lack of an in deep SQL knowledge. A.J.Millan ----- Original Message ----- From: "Igor Tandetnik" <[email protected]> To: <[email protected]> Sent: Monday, November 16, 2009 3:29 PM Subject: Re: [sqlite] Fine-grainy error report needed > A.J.Millan wrote: >> Supposing a dBase without constraints (some like this but not limited to >> it): >> >> CREATE TABLE tbl (Id INTEGER PRIMARY KEY, Nm INTEGER, Cl INTEGER, Dc >> REAL, >> Dm REAL, St INTEGER); >> >> A query to poblate: >> >> "INSERT INTO tbl (id,nm,cl,dc,dm,st) VALUES (?,?,?,?,?,?)" >> >> Then six binds (i = 0 to 5): >> >> sqlite3_bind_xxx (pStmt, i, ...); >> >> and a "step": resp = sqlite3_step(pStmt);. >> >> Then, if resp == SQLITE_MISMATCH (Data type mismatch) >> >> Do is there some way to know the index (i) of the offending bind? > > It's not quite clear what you expect SQLite to report, in general. > Consider: > > insert into tbl(id) values (? + ?); > insert into tbl(id) values (case when ? then ? else ? end); > > Which bind parameter is the offending one? > > Igor Tandetnik > _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
