[wktraster] ST_PixelSizeY sometimes gives negative answers

[robe]

Started to integrate WKT Raster in doco and came across this strange thing: where I get -0.5 for pixel size y

​http://www.postgis.org/documentation/manual-svn/RT_ST_PixelSizeY.html

This is using this which are excerpted from regress examples

​http://www.postgis.org/documentation/manual-svn/RT_reference.html

[Mike Taves]

Granted I don't know much about wktraster, this seems normal to me since a Y-pixel size is a unit vector, not an absolute value. The negative implies that that increments in Y-pixels are in the opposite direction to the Northings, which is normal since rasters have an origin in the upper-left corner and increment downwards while northings increment upwards—therefore a negative vector is required.

This metric in an ESRI World File is typically negative; see line 4:

• ​http://en.wikipedia.org/wiki/World_file
• ​http://resources.arcgis.com/content/kbase?fa=articleShow&d=17489

[robe]

That makes sense to me. I'm closing this off as a non-issue then.