Coordinates outside bounds of area of use now return inf

Running the GeoPandas tests with proj 6 / pyproj master, turned up this regression:

In [1]: import pyproj 

In [2]: pyproj.__version__ 
Out[2]: '2.1.3'

In [3]: p1 = pyproj.Proj(init='epsg:26918')
   ...: p2 = pyproj.Proj(init='epsg:4326')

In [4]: pyproj.transform(p1, p2, 0, 0)
Out[4]: (inf, inf)

vs

In [1]: import pyproj

In [2]: pyproj.__version__    
Out[2]: '1.9.6'

In [3]: p1 = pyproj.Proj(init='epsg:26918') 
   ...: p2 = pyproj.Proj(init='epsg:4326')  

In [4]: pyproj.transform(p1, p2, 0, 0)
Out[4]: (-79.48874388438705, 0.0)

So the transform now returns inf instead of an actual value. The coordinates however are out of the area of use (see bounds here http://epsg.io/26918). But previously it did calculate the transformed values nevertheless.

Here, it was a dummy test case in geopandas where we did not check that the tested dummy values were actually properly inside the bounds of the tested CRS. So that is easy to edit. But, it still is a breaking change (I don’t know that what extent people run into this in practice).

This seems a bit similar to https://github.com/pyproj4/pyproj/issues/202, but the difference is that here pyproj 1.9.6 actually returned values and not inf.

[snowman2]

@yxdragon, the good news is that there is a simple solution to your problem. In PROJ 6+, they take the axis order into consideration when doing transformations.

In your case, wkt1 has the latitude first longitude second and wkt2 has the east first and the northing second:

>>> from pyproj import CRS, Transformer
>>> wkt2 = 'PROJCS["China_Lambert_Conformal_Conic",GEOGCS["GCS_Beijing_1954",DATUM["Beijing_1954",SPHEROID["Krassowsky_1940",6378245,298.3,AUTHORITY["EPSG","7024"]],AUTHORITY["EPSG","6214"]],PRIMEM["Greenwich",0],UNIT["degree",0.0174532925199433]],PROJECTION["Lambert_Conformal_Conic_2SP"],PARAMETER["standard_parallel_1",30],PARAMETER["standard_parallel_2",62],PARAMETER["latitude_of_origin",0],PARAMETER["central_meridian",105],PARAMETER["false_easting",0],PARAMETER["false_northing",0],UNIT["metre",1,AUTHORITY["EPSG","9001"]]]'
>>> wkt1 = 'GEOGCS["WGS 84",DATUM["WGS_1984",SPHEROID["WGS 84",6378137,298.257223563,AUTHORITY["EPSG","7030"]],AUTHORITY["EPSG","6326"]],PRIMEM["Greenwich",0],UNIT["degree",0.0174532925199433],AUTHORITY["EPSG","4326"]]'
>>> crs2 = CRS(wkt2)
>>> crs2
<Projected CRS: PROJCS["China_Lambert_Conformal_Conic",GEOGCS["GCS ...>
Name: China_Lambert_Conformal_Conic
Axis Info [cartesian]:
- E[east]: Easting (metre)
- N[north]: Northing (metre)
Area of Use:
- undefined
Coordinate Operation:
- name: unnamed
- method: Lambert Conic Conformal (2SP)
Datum: Beijing 1954
- Ellipsoid: Krassowsky_1940
- Prime Meridian: Greenwich

>>> crs1 = CRS(wkt1)
>>> crs1
<Geographic 2D CRS: EPSG:4326>
Name: WGS 84
Axis Info [ellipsoidal]:
- Lat[north]: Geodetic latitude (degree)
- Lon[east]: Geodetic longitude (degree)
Area of Use:
- undefined
Datum: World Geodetic System 1984
- Ellipsoid: WGS 84
- Prime Meridian: Greenwich

Since this is the case, you need to swap the order of your input so that the latitude is first and longitude is second:

>>> transformer = Transformer.from_crs(crs1, crs2)
>>> transformer.transform(40, 100.20201515)
(-396207.328443438, 4808545.278686934)

Or, if you want it to always have the longitude first and latitude second, you can toggle on the always_xy argument:

>>> transformer = Transformer.from_crs(crs1, crs2, always_xy=True)
>>> transformer.transform(100.20201515, 40)
(-396207.328443438, 4808545.278686934)

For more information about this, I would recommend looking at:

• https://pyproj4.github.io/pyproj/stable/examples.html#transformations-from-crs-to-crs
• https://proj.org/faq.html#why-is-the-axis-ordering-in-proj-not-consistent

Hopefully this helps.

[snowman2]

It would be useful to have in another issue so it is easier for someone who has a similar question. I would recommend looking into Transformer.from_pipeline().