Lambert cylindrical projection support in PostGis
Does PostGis support Lambert cylindrical projection or is there a way to add support for it? I'm probably searching the wrong terms, but google isn't bringing up anything useful. Any help is appreciated.
EDIT: if not Lambert, are there are other cylindrical equal-area projections which can be used in PostGis.
It's built in. POSTGis uses the Proj.4 library for projections. The source code for Lambert equal area is PJ_cea.c (in version 4.7.0); a better-documented version appears as cea.js in Proj4js version 1.0.2. Here is the header from the .js version:
NAME LAMBERT CYLINDRICAL EQUAL AREA PURPOSE: Transforms input longitude and latitude to Easting and Northing for the Lambert Cylindrical Equal Area projection. This class of projection includes the Behrmann and Gall-Peters Projections. The longitude and latitude must be in radians. The Easting and Northing values will be returned in meters.
Let's get back to your original problem, because we may have lost sight of it during the intervening questions:
Basically I'm generating a grid table. Then I'll be doing a join query on some points and the grid table and group them by the cell they fall into to create histograms.
You can do this easily without generating any grid at all, because the formulas for an equal-area rectangular grid in decimal degrees are simple:
Let the intended width be w and the desired area be A. On the plane, a rectangle's height would nominally be h = A / w. Once and for all, convert w and h from meters to decimal degrees, using a conversion of approximately 10,000,000/90 meters per degree. Suppose the grid's origin (lower left corner) is at (x0,y0) as longitude and latitude. To determine what grid cell a point with (lon,lat) equal to (x,y) would fall in,
Compute x-x0 (as decimal degrees) and divide this by w. Round the result up to the nearest integer: this is the grid column index, j.
Compute sin(y)-sin(y0) and divide by h. Round up to the nearest integer: this is the grid row index, i.
The ordered pair (i,j) identifies the grid cell: use this to bin the points in the histogram. It won't enable you to draw a picture of the grid, but maybe you don't need that for your analysis…