On 11/20/2013 01:36 PM, Tom Lane wrote:
>
> You'd have to make the data self-identifying (which I think was the plan
> already), and ensure that *every* function taking "json" could cope with
> both formats on input. The typmod could only be expected to be enforced
> when storing or explicitly casting to one subformat, much like operations
> on "numeric" pay little attention to the original precision/scale if any.
>
> I agree that this solution isn't terribly workable, mainly because it'd
> break any third-party C functions that take json today.
>
>
Yeah, I had come to this conclusion. I don't think we can bolt on
typmods after the event.
I don't think having a separate "jsonb" type will be a tragedy.
I'm already planning on overloading the existing json functions and
operators.
cheers
andrew