[postgis-devel] Z() and M() semantics

[Markus Schaber]

Hi, Strk,

strk at refractions.net wrote:
> M() is already in the 1.1 branch.
> 
> BTW, what about making Z() and M() return NULL
> rather then 0 for points w/out the requested dimension ?
> 
> I think NULL is the correct semantic, but nobody answered in September:
> 
> http://postgis.refractions.net/pipermail/postgis-devel/2005-September/001577.html

I agree that NULL is the correct semantic, and the old semantic can be
easily emulated using coalesce(M(point),0) (which is backwards
compatible with the old semantic, btw).

Markus