On 01/15/2014 11:25 AM, Andrew Dunstan wrote:
>
>
> But here is the latest patch from that.
And one more.
here is the current state of the jsonb part of this:
all but two sets of functions (jsonb_extract_path* and the associated
operators #> and #>>, and
jsonb_populate_record/jsonb_populate_record_set) have been reimplemented
using the faster processing of jsonb. Those two sets still fall back on
processing json text. These will be fixed very shortly. I'm waiting on
some promised documentation. There is also some code cleanup (add more
comments, use ereport() instead of elog() and the like) that needs to be
done. However, it all works, and is just pretty much completely
consistent with the way json works.
I'm sure Oleg, Teodor and I will be able to get this whole thing whipped
into shape pretty quickly.
cheers
andrew