On Mar 13, 2013, at 10:45 AM, Andrew Dunstan <and...@dunslane.net> wrote:

> If someone wants functions to enforce a stricter validity check (e.g. via a 
> check constraint on a domain), or to convert json to a canonical version 
> which strips out prior keys of the same name and their associated values, 
> then these should be relatively simple to implement given the parser API in 
> the current patch. But they aren't part of the current patch, and I think 
> it's way too late to be adding such things.

I think it might be good to get something like this into core eventually, 
otherwise I suspect that there will be a different version of it for every 
JSON-using project out there. And my first cut at it won’t descend into 
sub-objects.

> I have been persuaded by arguments made upthread that the best thing to do is 
> exactly what other well known json-accepting implementations do (e.g. V8), 
> which is to accept json with duplicate keys and to treat the later key/value 
> as overriding the former key/value. If I'd done that from the start nobody 
> would now be talking about this at all.

That’s true, though I might have started thinking about a canonicalizing 
function before long. :-)

Best,

David



-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Reply via email to