(GeometryTypeError: Unknown geometry type: featurecollection) How to extend a function to all values in a column dependent on another dataframe?

Question

I have written a function to check if a point is within a polygon, and if it is to return "True,{branch_name}" where branch_name is the name of the polygon, and when trying to extend this function to the whole column, I keep encountering the "GeometryTypeError: Unknown geometry type: featurecollection" error.

The function I have written is:

def latlon_check(intsct, area, branch): #intsct = point, area = polygon, branch= branch name
    check = intsct.within(shape(area))
    if check == True:
        within.append(f"True,{branch}")
    else:
        within.append(f"False")
    return within

intsct is in df1 - a dataframe with a few hundred rows in it

area and branch are in df2 - a data frame with 10 rows

Note: The function works fine when inputting single values as arguments.

I want to make a new column in df1 where every row will either say "False" or "True,{branch_name}" showing which branch the point is in.

Using:

df1['within'] = df1['intsct'].apply(latlon_check, args = (df2['area'],df2['branch']))

and get the error:

GeometryTypeError: Unknown geometry type: featurecollection

I have tried rewriting the column as a string then converting it back to 'geometry' and still had the same error, will appreciate any help!

Answer 1

A few points

• have used geopandas sample geometry to simulate data you describe. Making columns of two data frames consistent with what your describe
• your sample code uses a function shape() it's not clear what this is. Possibly takes a WKT string and converts it into a polygon. Have removed as a polygon is being passed in context I have simulated
• core you are passing series from df2 by the way you are calling apply(). Hence have refactored latlon_check() to use series rather than singletons.
• within.append() this is function to add items to a list there is no list. Have just simplified to return the desired string

MWE

import geopandas as gpd

# simulate data that matches question
df1 = gpd.read_file(gpd.datasets.get_path('naturalearth_cities'))
df2 = gpd.read_file(gpd.datasets.get_path('naturalearth_lowres'))
df1 = df1.rename(columns={"geometry":"intsct"})
df2 = df2.sample(10, random_state=42).rename(columns={"geometry":"area","name":"branch"}).loc[:,["branch","area"]]

def latlon_check(intsct, area, branch): #intsct = point, area = polygon series, branch= branch name series
    for i, area_ in area.items():
        if intsct.within(area_):
            return f"True,{branch.loc[i]}"
    return "False"
df1["within"] = df1['intsct'].apply(latlon_check, args = (df2['area'],df2['branch']))

sample output

	name	intsct	within
0	Vatican City	POINT (12.4533865 41.9032822)	False
1	San Marino	POINT (12.4417702 43.9360958)	False
2	Vaduz	POINT (9.5166695 47.1337238)	False
3	Lobamba	POINT (31.1999971 -26.4666675)	False
4	Luxembourg	POINT (6.1300028 49.6116604)	False
22	Podgorica	POINT (19.2663069 42.4659725)	True,Montenegro
57	Port-au-Prince	POINT (-72.3379804 18.5429705)	True,Haiti
84	Riga	POINT (24.0999654 56.9500238)	True,Latvia
114	Sucre	POINT (-65.2595156 -19.0409708)	True,Bolivia
119	Yerevan	POINT (44.5116055 40.1830966)	True,Armenia
122	La Paz	POINT (-68.151931 -16.4960278)	True,Bolivia
198	Ürümqi	POINT (87.5730598 43.8069581)	True,China
199	Chengdu	POINT (104.0680736 30.6719459)	True,China
214	Taipei	POINT (121.5683333 25.0358333)	True,Taiwan
227	Beijing	POINT (116.39420089260611 39.901720309862675)	True,China
232	Shanghai	POINT (121.4345588 31.2183983)	True,China
242	Hong Kong	POINT (114.1830635 22.3069268)	True,China