On Fri, 2 May 2014 21:09:46 +0200
Petite Abeille <petite.abei...@gmail.com> wrote:
>
> On May 2, 2014, at 8:54 PM, Richard Hipp <d...@sqlite.org> wrote:
>
> > I'm guessing that Mr. Abeille is upset that SQLite ?
>
> ? doesn?t even bother with SQL syntax and will happily accept any old
> junk as a sorry excuse for a query.
>
> select 1 where null;
> select 1 where not null;
>
> When SQLite 4 sees the light of the day, I wish for a strict SQL
> parser and proper error messages.
To amplify the point, the issue isn't pure fussiness or obligation to
adhere to standards. A permissive parser invites error.
It's not hard to imagine
select 1 where 1 - 1;
was intended as
select 1 where 1 = 1;
which, in the midst of a large query producing expected results, might
easily be overlooked.
I doubt Petite is confused by boolean evaluation, but rather is
dismayed by its appearance in this context.
SQL is not C. To the extent the SQL supplied by SQLite is nonstandard,
it might as well be another language entirely. The better one knows
SQL, the harder a nonconformant implementation is to use.
Not long ago I was helping someone with a query in MS Access. Easy,
just use a correlated subquery in an update statement. Hard, if
it chokes the parser. Perhaps you know the joke with the punchline,
"Assume a can opener."
--jkl
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
