如何找到连续的lat，lon点距离

Question

I have a df with a series of consecutive geo coordinates. I want to find the distance between these consecutive points. 1->2, 2->3 .... end-1->end.

Using df.shift(1) doesn't look pretty, using a loop either.

Can it be done more elegantly with some recursive functions?

Answer 1

The solution

import pandas as pd

def calculate_distance(lat_from, long_from, lat_to, long_to):
    # some better logic
    return lat_from - lat_to + long_from - long_to

df = pd.DataFrame({'long': [1, 2, 4.2, 5, 6], 'lat': [7, 4, 2, 1.2, 2]})
df[['lat_to', 'long_to']] = df.shift(-1)

# this is way faster, but may not be possible depending on your calculation
calculate_distance(df['lat'], df['long'], df['lat_to'], df['long_to'])
>>> 0    2.000000e+00
>>> 1   -2.000000e-01
>>> 2    2.220446e-16
>>> 3   -1.800000e+00
>>> 4             NaN
>>> dtype: float64

# or
# a lot slower, processes on per-row basis
df.apply(lambda row: calculate_distance(row['lat'], row['long'], row['lat_to'], row['long_to']), axis=1)
>>> 0    2.000000e+00
>>> 1   -2.000000e-01
>>> 2    2.220446e-16
>>> 3   -1.800000e+00
>>> 4             NaN
>>> dtype: float64

For speed comparison, try the difference between pandas.DataFrame.apply , pandas.DataFrame.applymap and normal broadcasted operations.