[postgis-users] intersection doesn't seem to intersect back w/ original geom

Sandro Santilli strk at keybit.net
Tue Jan 22 06:35:11 PST 2013


On Tue, Jan 22, 2013 at 09:24:18AM -0500, Charlton Galvarino wrote:
> Hi, all.  I'm completely stumped and hope the answer is right in front of my face.
> 
> The pink ls.
> a.geom = 0102000020BA08000002000000ED7E57CC2CDF2F41BCE00145EF9F29414DABCE8596DE2F41F40B15A36C9F2941
> 
> The blue ls.
> b.geom = 0105000020BA0800000100000001020000000A0000003E35FDD6D4DE2F413858F60E92A12941003235ED84DE2F410075ECAFEF9F294100F99BA4C2DE2F41000B8A37399E294180E6B00C42DF2F410049DDE92E9C2941805D5940DDDF2F41009D30FCC79A2941004F9FEB5CE02F41808734555298294100D58EBEB6E02F41005398EA9B9629410099A09767E02F4100CE6014A99529418024290838DF2F41008CA543DF932941801DD29CF2DD2F41804F933046932941
> 
> Their intersection (yellow dot) computed from the above two geoms.
> c.geom = 0101000020BA08000075E19F4797DE2F415AA1904B6D9F2941
> 
> My SRID's are all the same.
> 
> Why in the **** doesn't the final intersection ST_Intersects() w/ either original linestring?!  Precision issue?  I'm nuts issue?

Precision issue. What you can do is _snap_ the input lines against the
intersection points. Then you'll have it.

--strk;


More information about the postgis-users mailing list