[postgis-devel] I want boxes

Paul Ramsey pramsey at opengeo.org
Wed Nov 9 10:04:09 PST 2011


Any suggestions on the WKT form for a box?

EXTENT(minx, miny, max, maxy) ?
EXTENT(minx, maxx, miny, maxy) ?

the second form is what I actually store in serialization because
keeping the dimensions together makes it easier to have more/less
dimensions. However, I think the first form is better from a usage
pint of view... probably would be better to follow some WKT strictures
and embed POINT in the form like this

EXTENT(minx miny, maxx maxy)

And then the extended forms become

EXTENT Z (minx miny minz, maxx maxy maxz)
EXTENT M (minx miny minm, maxx maxy maxm)
EXTENT ZM (minx miny minz minm, maxx maxy maxz maxm)

Any thoughts on the type token? I feel a little "box overload" so am
suggesting "extent". But it could be box.

Let me know, this is supposed to be my item before freeze...

P.


On Tue, Nov 8, 2011 at 12:19 PM, Sandro Santilli <strk at keybit.net> wrote:
> On Tue, Nov 08, 2011 at 10:49:00AM -0800, Chris Hodgson wrote:
>>
>> >>If I'm right, perhaps a non-agreggate st_extent() function is in order, because that doesn't seem like such a weird thing to do yet requires a lot of function calls to do efficiently.
>> >
>> >Note that ST_Extent drops dimensionality and SRID.
>>
>> True. I suppose what you are doing is unusual enough that we don't
>> care if the right way to do it is a little esoteric. It just makes
>> me wonder when a core developer has to ask on the list if they are
>> doing something so relatively simple, correctly.
>
> Well, I actually belive what I'm doing is pretty common: maintain
> an envelope and growing it with the partecipation of new geometries.
> It _can_ be done, only it sounds more complex once you want to avoid
> using the good old "box2d" type. As the subject suggest this was more
> about sharing a use case for a "box" type.
>
> --strk;
>
>  ()   Free GIS & Flash consultant/developer
>  /\   http://strk.keybit.net/services.html
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>



More information about the postgis-devel mailing list