On Sun, Nov 21, 2010 at 12:31 AM, Terry Laurenzo <tj@laurenzo.org> wrote:
> What we see from this is that parsing JSON text and generating a binary
> representation is cheap, representing approximately 10% of the base case
> time. Conversely, anything that involves generating JSON text is expensive,
> accounting for 30-40% of the base case time. Some incidental profiling
> shows that while the entire operation is expensive, the process of
> generating string literals dominates this time. There is likely room for
> optimization in this method, but it should be noted that most of these
> documents are lightly escaped (if escaped at all) which represents the happy
> path through the string literal output function.
Ouch! That's kind of painful. But certainly for some use cases it
will work out to a huge speedup, if you're doing subscripting or
similar.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company