Empirically, I came up with 0.0000104 degrees/meter for my neck of the woods.


On Tue, May 28, 2013 at 6:23 PM, Charalampos Nikolaou <charnik@di.uoa.gr> wrote:
Hi,

On Tue, May 28, 2013 at 9:13 PM, Casey Gum <casey.gum@gmail.com> wrote:
> In the docs about the geospatial extension, it mentions a buffer function:
>
> CREATE FUNCTION Buffer(a Geometry, distance FLOAT) RETURNS Geometry
>
> Does anyone know what a good value would be to use for distance if I wanted
> to buffer a point, line, etc. X number of meters?  Or does anyone have any
> ideas for what would be a good approach for this problem - where you want
> the user to be able to specify how close two things can be to one another to
> be a match?

Actually MonetDB assumes that all geometry objects are internally
represented in a specific spatial reference system and in particular
EPSG 4326 (http://spatialreference.org/ref/epsg/4326/) aka WGS84.
Since the units of measurement for such a reference system is degrees,
then whenever you need to specify a distance value, you have to
specify it in degrees. If I am not mistaken, a good value for a
distance of 2 to 10 kilometers might be around to 0.002 degrees or
something like that, but take this with a grain of salt. Generally
speaking, there is no single answer and 1-1 transformation of meters
to degrees since in different geographical areas this correspondence
variates (for example by moving from the equator to a pole). And this
has to do with the ellipsoid approximation of the Earth made by WGS84.
See for example this wikipedia entry
http://en.wikipedia.org/wiki/Decimal_degrees. Thus, if you are
interested in a specific area, you might be able to "tune" that value.

Hope that helps,
Babis

>
> Thanks a lot,
>
> Casey
>
> _______________________________________________
> developers-list mailing list
> developers-list@monetdb.org
> http://mail.monetdb.org/mailman/listinfo/developers-list
>
_______________________________________________
developers-list mailing list
developers-list@monetdb.org
http://mail.monetdb.org/mailman/listinfo/developers-list