Darren Duncan wrote:
> Roger Binns wrote:
>
>>> In fact this support might even be easier as it may only require
>>> enhancements to
>>> the SQL parser, which would generate VM opcodes like for a CHECK
>>> constraint,
>>> unless further work is done to optimize for the presented cases, or
On Fri, Oct 30, 2009 at 03:59:11PM -0500, Jay A. Kreibich wrote:
> On Fri, Oct 30, 2009 at 03:19:59PM -0500, Nicolas Williams scratched on the
> wall:
> > I should add that a pragma that cause CHECK constraints to be
> > automatically created for enforcing strong typing in subsequent CREATE
> >
On Fri, Oct 30, 2009 at 03:19:59PM -0500, Nicolas Williams scratched on the
wall:
> I should add that a pragma that cause CHECK constraints to be
> automatically created for enforcing strong typing in subsequent CREATE
> TABLE statements
That's tricky. Values have TYPES. Columns have
I think that something several people had missed was that I specifically
proposed the strong typing behavior to be activated by a new pragma, and unless
people activate that pragma they would get the old behavior, so total backwards
compatibility. I see several people then proposed using the
I should add that a pragma that cause CHECK constraints to be
automatically created for enforcing strong typing in subsequent CREATE
TABLE statements is rather like having FOREIGN KEY clauses automatically
generate triggers. There's precedent, in other words, and it is a
simple way to implement
On Fri, Oct 30, 2009 at 01:30:31PM -0400, John Crenshaw wrote:
+1
I don't think this proposal can or will be accepted.
One reasonable idea, perhaps, would to have a pragma that causes
subsequent CREATE TABLE statements to get automatically generated CHECK
expressions that enforce typing. Any
> I believe I understand Darren's point (whether or not I care for them
> is another story).
Yes, you've understood Darren for the most part, but clearly don't understand
the objections.
> On Fri, Oct 30, 2009 at 2:22 AM, Roger Binns wrote:
> > -BEGIN PGP SIGNED
On 30 Oct 2009, at 2:14pm, P Kishor wrote:
> Actually, there can be one bad effect of Darren's suggestion, now that
> I think of it, and that would be for those who don't care for strong
> typing. They will end up getting strong typing for all non-UNIVERSAL
> columns whether they like it or not,
On Fri, Oct 30, 2009 at 7:14 AM, P Kishor wrote:
> Actually, there can be one bad effect of Darren's suggestion, now that
> I think of it, and that would be for those who don't care for strong
> typing. They will end up getting strong typing for all non-UNIVERSAL
> columns
On Oct 30, 2009, at 10:14 AM, P Kishor wrote:
> Actually, there can be one bad effect of Darren's suggestion, now that
> I think of it, and that would be for those who don't care for strong
> typing. They will end up getting strong typing for all non-UNIVERSAL
> columns whether they like it or
I believe I understand Darren's point (whether or not I care for them
is another story).
On Fri, Oct 30, 2009 at 2:22 AM, Roger Binns wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Darren Duncan wrote:
>> But on a newer SQLite that implements the stronger
Been watching this discussion go back and forth, and I'd like to weigh
in.
I'm generally a HUGE fan of strong typing, but this doesn't do it for
me. To me, strongly typed means a compiler catches my type mismatches
before the app goes out the door. In this case though, no matter what
you do, a
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Darren Duncan wrote:
> But on a newer SQLite that implements the stronger typing support I proposed,
> when that feature is active then columns with declared types like INTEGER/etc
> would enforce that only values of that type are stored there,
I
Roger Binns wrote:
>> In fact this support might even be easier as it may only require
>> enhancements to
>> the SQL parser, which would generate VM opcodes like for a CHECK constraint,
>> unless further work is done to optimize for the presented cases, or to
>> enhance
>> semantics.
>
> It
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Darren Duncan wrote:
> I should point out, for people reading this thread, that it is very possible
> to
> have both the amount of type flexibility that SQLite provides and have strict
> typing, at the same time.
What I fail to understand with all
This email is a feature proposal for SQLite; I can rewrite it if desired.
Ted Rolle wrote (in the "Late data typing ..." thread):
> Doesn't dynamic data typing lead to bad data?
> And proliferation of home-grown editing routines?
> It seems that a strict data typing at column definition time
16 matches
Mail list logo