Trouble with using GDAL SetConfigOption in Jupyter Notebook running on Docker

Question

I am trying to use GDAL in Jupyter Notebook which is running on the Docker environment, the GDAL library work fine but when it comes to set a configuration by using SetConfigOption it does not work.

I am using Use jupyter/minimal-notebook image in the Docker file and I did t install GDAL on the image as dependencies by using pip and conda but none of them seems to be responding to SetConfigOption, and when I am using the GDAL in Jupyter Notebook in Windows everything seems to be in order and nothing is wrong. Why does this happen?

(When I print the SetConfigOption it says that the configuration has been set but in reality it does not affect anything.)

This my Jupyter Notebook file:

# Activating the GDAL Driver  
# Set GDAL Config Parameters
# importing the libraries "GDAL" and "OGR"
# importing the library "os" with helps with some operating system specific i/o tasks (working    with paths, directories, files, ..) 

try:
    from osgeo import gdal, ogr
    import os 
except:
     print ('ERROR: cannot find GDAL/OGR modules')

 gdal.UseExceptions()

 gdal.SetConfigOption("CPL_LOG_ERRORS","ON")
 gdal.SetConfigOption("VALIDATE", "YES")
 gdal.SetConfigOption("FAIL_IF_VALIDATION_ERROR","YES")
 gdal.SetConfigOption('WRITE_GFS','NO')


 print(gdal.GetConfigOption("WRITE_GFS"))

 driverName="GMLAS"

 gml_driver = ogr.GetDriverByName(driverName)

 if gml_driver is None:
     raise Exception('GDAL driver >"+driverName+"< not available')
 else: print("GDAL driver successfully loaded: >"+gml_driver.GetName()+"<")


 gdal.VectorTranslate(
    'test.geojson',
    'Simple_XtraServerGetFeature (3).xml',
     options='-f "GeoJSON"'
 )
 print(gdal.GetConfigOption("WRITE_GFS"))

This my Docker file:

# Use the jupyter/minimal-notebook image as the parent image
FROM jupyter/minimal-notebook:latest

# Install required dependencies
RUN conda install -c conda-forge gdal
RUN pip install --no-cache-dir \
    jupyter \
    requests>=2.21.0
# Set the working directory to /src
WORKDIR /src

# Make port 8888 available to the world outside this container
EXPOSE 8888

# Define environment variable
ENV NAME World

# # Copy the current directory contents into the container at /app
# COPY . /src

# Mount the host machine directory into the container
VOLUME /src

# Run Jupyter Notebook when the container launches
CMD ["jupyter", "notebook", "--ip=0.0.0.0", "--port=8888", "--no-browser", "--allow-root"]

Answer 1

The issue is likely that you need to explicitly close the dataset returned by gdal.VectorTranslate():

ds = gdal.VectorTranslate(....)
del ds  # ogr ds.Close() with GDAL >= 3.8.0