[postgis-users] Dateline and polar regions and any other spatial gotchas

[Paul Ramsey]

There are definitely issues with dateline and poles when working in 
4326, since the coordinates are treated as planar, not polar. 
Workarounds for the dateline are relatively easy to do, usually in the 
form of double-entry inserts (we should probably write a standard 
trigger for doing that (if something crosses the dateline, insert it 
twice, once on each side of the plane)).  The poles I think are a harder 
problem.

P

Gregory Williamson wrote:
> Most of the work our company (GlobeXplorer) has done with postGIS has 
> been in US markets where such things as dateline math isn't much of an 
> issue. We use SRID 4326 to store the data in.
> 
> We (DigitalGlobe) are evaluating PostgreSQL and Oracle for a largish 
> project using data all over the planet, and before I make any bald 
> assertions I thought I'd ask and see if there are:
> 
> a) know issues with datelines or other gnarly parts of the world ?
> 
> b) workarounds ?
> 
> Thanks for any info ... tried trolling the website and didn't find any 
> caveats, but the mail list archive isn't really searchable so I am 
> hoping this is not a FAQ (if it is, point me at it!).


-- 

   Paul Ramsey
   Refractions Research
   http://www.refractions.net
   pramsey at refractions.net
   Phone: 250-383-3022
   Cell: 250-885-0632