Calling GDAL/OGR commands from Python

key words: python, GDAL/OGR, subprocess.check_call, os.walk, external commands, environment variables

If you plan to call GDAL/OGR commands from python code you might need to set up environment variables before you can do that. Environment variables can be defined at system level and in that case you don’t need to do that but when they are not defined on the system level you will have to take care of them.

Calling external commands from python can be accomplished by subprocess module. There are several function that can be useful for this task but it this post I didn’t show all the differences between these function the main thing I wanted to show you is how to set up environment variables required by some external program called by subprocess module. In this particular case external program is any of GDAL/OGR commands.

If you want to write porogram that calls GDAL/OGR I can assume that you need it to process multiple files and that is why this sample program uses os.walk function. Walk function allows you to process multiple files in the multiple folders. This sample filters all tif files that can be found in grid_data folder.

Using subprocess subprocess.check_call function I have demonstrated invocation of gdalinfo command for all TIFF files and just as an example one OGR command (ogrinfo). The same principles can be applied for file processing using any of available GDAL/OGR commands.

Here is the source code.

.... enjoy ...

'''
Created on 10. okt. 2012
@author: Stipica Pavicic
'''
import os, fnmatch, subprocess

for (root, dirs, files) in os.walk(r'L:\sp\python\grid_data'):
    #print (root)    
    #print (os.listdir(root)) # list of files and folders
    #print (fnmatch.filter(os.listdir(root),".")) # filter the list of files and folders
    
    for filename in fnmatch.filter(files,"*.tif"):
      
        fbasename, extension = os.path.splitext(filename)
        
        gdal_root = r'C:\release-1310-gdal-1-9-mapserver-6-0'
        pathv=''
        pathv= pathv + gdal_root  + r'\bin;' 
        pathv= pathv + gdal_root  + r'\bin\gdal\python\osgeo;' 
        pathv= pathv + gdal_root  + r'\bin\proj\apps;' 
        pathv= pathv + gdal_root  + r'\bin\gdal\apps;' 
        pathv= pathv + gdal_root  + r'\bin\ms\apps;' 
        pathv= pathv + gdal_root  + r'\bin\gdal\csharp;'  
        pathv= pathv + gdal_root  + r'\bin\ms\csharp;' 
        pathv= pathv + gdal_root  + r'\bin\curl;'
        pathv= pathv + gdal_root  + r'\bin\gdal\apps'      
        
        
        #if you don't work with Oracle or you don't need Oracle plugin just rename/remove \bin\gdal\plugins\gdal_GEOR.dll & ogr_OCI.dll files to avoid unnecessary error messages
        env_cust={"GDAL_DATA":        gdal_root + r'\bin\gdal-data', \
                  "GDAL_DRIVER_PATH": gdal_root + r'\bin\gdal\plugins',  \
                  "SET PYTHONPATH":   gdal_root + r'\bin\gdal\python\osgeo;' +  
                                      gdal_root + r'\bin\ms\python', \
                  "SET PROJ_LIB":     gdal_root + r'\bin\proj\SHARE', \
                  "PATH": pathv}
        
        os.chdir(root)
        #print (root, filename, fbasename, extension)
        #subprocess.check_call(['echo' , filename], env=env_cust, shell=True)
        subprocess.check_call(['gdalinfo', filename], env=env_cust, shell=True)
        subprocess.check_call(['ogrinfo', '--help-general'], env=env_cust, shell=True)