Open GDB file in Python

Question

I would like to be able to open and view a GDB file (GIS file for raster data) in Python.

The data is official US government GIS data, the file can be downloaded publicly here: https://www.fs.usda.gov/detail/r5/communityforests/?cid=fseprd1009698

there is also an arcgis app that hosts the data so that it can be viewed: https://usfs.maps.arcgis.com/apps/webappviewer/index.html?id=d1bb42e8e14c46528bbedcede6b03e62

When clicking the download button a zip file is downloaded that has a folder with many GIS files within it.

I am looking into two options:

Is there a way to directly read a GDB folder or zip file in Python?

Is there a way to convert the GDB file to another format like GeoTIFF and then load that file into Python?

Answer 1

I downloaded the Oahu dataset and had a try with GDAL v3.8.4 installed with OSGeo4W.

GDAL can read the raster layer but it is awfully slow to read the whole dataset. It may be better to convert the data into GeoTIFF and add overviews. However, it seems that the mask band that GDAL creates makes even the GeoTIFF version very slow to view with QGIS at small scale. That happens because GDAL does not create automatically overviews for the .msk band. The fix is to create overviews for .msk manually. GDAL 3.9 will do it automatically.

gdal_translate -of gtiff -co tiled=yes -co compress=deflate Oahu_Canopy_2021.gdb oahu_canopy_2021.tif
gdaladdo oahu_canopy_2021.tif
gdaladdo oahu_canopy_2021.tif.msk

The FileGDB version does not have overviews and it is usable for browsing only at scales 1:20000 and larger.

What I tested with gdal_translate applies also to GDAL Python bindinds. It is possible to read data directly from the FileGDB but it may make sense to convert data into GeoTIFF first.

• user2856: @MustardTiger The rendering can be improved with overviews gdaladdo -ro --config COMPRESS_OVERVIEW DEFLATE --config NBITS 1 oahu_canopy_2021.tif 2 4 8 16. Notes: --config NBITS 1 is required as the raster is 1 bit; default resampling is nearest but can be changed using -r [resampling_method] e.g. -r average, see gdaladdo documentation.
• user30184: At least my GDAL version keeps NBITS as 1 without specially demanding it. Gdaladdo was not enough for making rendering fast because rendering was still slow due to the full resolution .msk file. Therefore the need to run gdaladdo also for the .msk file. In the next GDAL version that will happen automatically github.com/OSGeo/gdal/pull/9603. Compressing the overviews is a good suggestion.
• user2856: "my GDAL version keeps NBITS as 1 without specially demanding it", yes I'm also running v3.8.4, but my NBITs 1 comment was referring to gdaladdo which errored if I didn't specify --config NBITS 1 not gdal_translate which automatically output a 1 bit tiff from the FGDB raster.
• user30184: gdaladdo -ro --config COMPRESS_OVERVIEW DEFLATE oahu_canopy_2021.tif works for me. What error did you get?

Answer 2

A file geodatabase can store both vector and raster data, see What is a file geodatabase?.

Vector data can be read using geopandas

import geopandas as gpd
import fiona
file_geodatabase = r"C:\GIS\GIStest\test.gdb" #A file geodatabase with multiple vector layers in it
#Print them
for feature_class in fiona.listlayers(file_geodatabase):
 print(feature_class)
 # test
 # randompoints
 # roads_with_gap_GenerateNearT
 # roads_with_gap_G_TableSelect
 # nature
df = gpd.read_file(filename=file_geodatabase, layer="randompoints") #Read one layer
print(df.head(2))
# ID Value Date geometry
# 0 A 1 2023年05月31日 00:00:00+00:00 POINT (714582.638 6728435.051)
# 1 A 5 2023年06月01日 00:00:00+00:00 POINT (1149955.918 6240796.036)
df.to_file(r"C:\GIS\GIStest\random_points_converted.gpkg") #Save as geopackage

If you have raster layers you should be able to use GDAL

• Mustard Tiger: I ran your code and got the following error: not recognized as a supported file format.
• Bera: Is it a raster or vector layer you are trying to open?
• user2856: @MustardTiger You need a recent version of GDAL. Raster support was added to the FileGDB driver in GDAL 3.7.