Preserve column order of GeoPandas file read

Question

I have a shapefile I'm reading in using GeoPandas and then I do some geometry conversion of the data but in the process it unfortunately sorts the columns alphabetically. I need to preserve original order of the columns which has over 200 columns.

Here is the start of the code I'm using. I imagine I have to somehow store the original column order and apply it to the GeoPandas dataframe.

import geopandas as gpd


shpfile = r"C:\data\myfile.shp"
poly = gpd.read_file(shpfile)
poly.head()

Answer 1

If you just need a variable that stores the index at the end of those 4 lines for later use and then uses that to sort the dataframe down the line, then how about:

poly['stored_index'] = poly.index
...
poly.sort_values('stored_index', inplace=True)

Answer 2

store the column order right after read-in...

import geopandas as gpd

shpfile = r"C:\data\myfile.shp"
poly = gpd.read_file(shpfile)
cols = poly.columns
# ...do something that will change the order of the poly GDF...
# ...then, as long as your work hasn't created any new columns...
poly = poly[cols]

if you add new columns do..

cols = poly.columns.tolist()

and then manipulate the new column/s into that list where you want them in the GDF.

Answer 3

What version of geopandas are you using? I found that when I read in a shapefile with v0.6.1, the column order was somewhat alphabetized without any command from my script to do so. When I ran the exact same code with the same version of pandas (v1.3.5) and geopandas upgraded to v0.13.2, the resulting dataset (in GeoJSON format) had the original column order from the input .shp preserved. I'm not sure what changed between these versions or when it changed (maybe v0.11?), but I'd recommend setting specific requirements for more recent versions of geopandas to fix this.