Advanced Examples

Optimize Transformations

Here are a few tricks to try out if you want to optimize your transformations.

Repeated transformations

If you use the same transform, using the pyproj.Transformer can help optimize your transformations.

import numpy as np
from pyproj import Transformer, transform

transformer = Transformer.from_proj(2263, 4326)
x_coords = np.random.randint(80000, 120000)
y_coords = np.random.randint(200000, 250000)

Example with transform():

transform(2263, 4326, x_coords, y_coords)

Results: 160 ms ± 3.68 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

Example with Transformer:

transformer.transform(x_coords, y_coords)

Results: 6.32 µs ± 49.7 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)

Tranforming with the same projections

pyproj will skip transformations if they are exacly the same by default. However, if you sometimes throw in the projections that are about the same and the results being close enough is what you want, the skip_equivalent option can help.

Note

From PROJ code: The objects are equivalent for the purpose of coordinate operations. They can differ by the name of their objects, identifiers, other metadata. Parameters may be expressed in different units, provided that the value is (with some tolerance) the same once expressed in a common unit.

Transformation Group

The TransformerGroup provides both available transformations as well as missing transformations.

  1. Helpful if you want to use an alternate transformation and have a good reason for it.

>>> from pyproj.transformer import TransformerGroup
>>> trans_group = TransformerGroup("epsg:4326","epsg:2964")
>>> trans_group.best_available
True
>>> trans_group.transformers[0].transform(66, -153)
(149661.2825058747, 5849322.174897663)
>>> trans_group.transformers[1].transform(66, -153)
(149672.928811047, 5849311.372139239)
>>> trans_group.transformers[2].transform(66, -153)
(149748.32734832275, 5849274.621409136)

  1. Helpful if want to check that the best possible transformation exists. And if not, how to get the missing grid.

>>> from pyproj.transformer import TransformerGroup
>>> tg = TransformerGroup("epsg:4326", "+proj=aea +lat_0=50 +lon_0=-154 +lat_1=55 +lat_2=65 +x_0=0 +y_0=0 +datum=NAD27 +no_defs +type=crs +units=m", always_xy=True)
UserWarning: Best transformation is not available due to missing Grid(short_name=ntv2_0.gsb, full_name=, package_name=proj-datumgrid-north-america, url=https://download.osgeo.org/proj/proj-datumgrid-north-america-latest.zip, direct_download=True, open_license=True, available=False)
"{!r}".format(operation.grids[0])
>>> tg.best_available
False
>>> tg.transformers[0].description
'axis order change (2D) + Inverse of NAD27 to WGS 84 (3) + axis order change (2D) + unknown'
>>> tg.unavailable_operations[0].name
'Inverse of NAD27 to WGS 84 (33) + axis order change (2D) + unknown'
>>> tg.unavailable_operations[0].grids[0].url
'https://download.osgeo.org/proj/proj-datumgrid-north-america-latest.zip'