On Mon, 3 Feb 2020 10:45:50 +0100
Dominique Devienne <ddevie...@gmail.com> wrote:
> On Sun, Feb 2, 2020 at 12:50 AM Richard Hipp <d...@sqlite.org> wrote:
>
> > On 2/1/20, Thomas Kurz <sqlite.2...@t-net.ruhr> wrote:
> > > Does this mean there will be no possibility to prevent inserting
> > > a string into an integer column anymore?
> > >
> > > create table x (x integer check (typeof(x) == 'integer'));
> > > insert into x values ('1');
> > >
> > > --> will pass in future versions???
> >
> > I think that is what it means. yes.
>
> Wow... I haven't caught up on this thread, but that's really really
> bad IMHO
...
> the fact we can no longer do that would be a real shame. I wonder
> where this is coming from... --DD
It's a good thing, really! The rule would be that the provided value is
converted to the column's type before inserting. If it can't be
converted, it's still an error. If it can, great. I think you'll find
that's the behavior of most SQL DBMS implementations.
After all, of what significance is the type of the provided argument?
Do you want to force applications to "pre-convert" values the DBMS can
convert implicitly? Do you want binding choices in the application to
drive the datatype in the database, or do you want the database to
enforce types?
--jkl
_______________________________________________
sqlite-users mailing list
sqlite-users@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
