gdal_edit in Python

Question

I am new of programming and I need help for it. I would use gdal_edit to modify SRC of 300 hundred aerial photos, if I use this tool one by one work perfectly, but I would want to use it in a cycle in a python script to simplify and get faster my job.

This is the tool I used:

gdal_edit -a_srs EPSG:3004 "C:/Users/Laptop/Desktop/test/img.ecw" 

While this is the python script in that I would use this tool:

import os, gdal
dir="C:/Users/Laptop/Desktop/test"
array=os.listdir(dir)
for e in array:
 if os.path.splitext(e)[1]==".ecw":
 print(e)
 gdal_edit -a_srs EPSG:3004 dir+e

How can I use gdal_edit in this code? I use OsGeo4W Shell that has gdal and python preinstallated.

• Jon: One way is to create a string that matches exactly what you would type normally, then use subprocess.call(string) to execute it. You will need to import subprocess module. Also, I have noticed that it is best to create your string as a list, i.e. string = ['gdal_edit', '-a_srs', 'EPSG:3004', 'C:/Users/Laptop/Desktop/test/img.ecw']
• Luca Caliciotti: Hi Jon, thank you for your post, I tried using your advice before in direct command: import subprocess subprocess.call['gdal_edit', '-a_srs', 'EPSG:3004', 'C:/Users/SIGEO/Documents/Luca/temp/test/521142.ecw'] but this is been error answer: Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: 'function' object has no attribute 'getitem' Can you understand it?
• Jon: If you're using the exact code you posted, you have to call subprocess.call with (), not with []. Break it up as I suggested in my first comment: string = ['gdal_edit'...etc] then subprocess.call(string). It's more readable, and, as you discovered, less prone to mistakes. If you want to use your one-liner, just wrap your list in (); i.e. subprocess.call(['gdal_edit', '-a_srs', 'EPSG:3004', 'C:/Users/SIGEO/Documents/Luca/temp/test/521142.ecw'])

Answer 1

The gdal_edit program is written in Python. You can copy a few lines of its code into your loop over the names of files to modify. This approach will be much faster than executing gdal_edit in a subprocess because creating new processes and launching Python takes a significant amount of time.

Here's how it opens the dataset for editing: https://github.com/OSGeo/gdal/blob/master/gdal/swig/python/scripts/gdal_edit.py#L280.

Here's how it updates the dataset's spatial reference system: https://github.com/OSGeo/gdal/blob/master/gdal/swig/python/scripts/gdal_edit.py#L310-L319.

• Mr Tarsa: Links are broken now, due to migration