How calculate euclidean distance between pandas and list?

Question

I have a pandas dataframe and I try to calculate all euclidean distance with a fixed value and find the shortest distance.

My dataframe "currency":

        Stype   h  line        ...            y    y2                bc
45   currency  38    13        ...         1344  1382  (1731.0, 1363.0)
46   currency  38    13        ...         1343  1381  (2015.0, 1362.0)
47   currency  39    13        ...         1342  1381  (2267.5, 1361.5)
60   currency  39    15        ...         2718  2757   (488.0, 2737.5)
61   currency  39    15        ...         2717  2756   (813.5, 2736.5)
62   currency  39    15        ...         2718  2757  (1332.5, 2737.5)
63   currency  40    15        ...         2716  2756  (1821.5, 2736.0)
64   currency  39    15        ...         2715  2754  (2286.5, 2734.5)
68   currency  39    17        ...         2874  2913  (2287.5, 2893.5)
162  currency  30    22        ...         3311  3341  (1104.5, 3326.0)

example value in my list [l['bc']]

[(2126.5, 2657.0)]

My code:

for l in label_dic:
    print('bc:', [l['bc']])
    print(cdist([l['bc']], currency.bc.values, 'euclidean'))

My issue:

ValueError: XB must be a 2-dimensional array.

I have validated my function with:

print(cdist([l['bc']], [l['bc']], 'euclidean'))
Result: [[0.]]

Can you exaplin me how to fix it ?

Thanks

Answer 1

currency.bc.values seems to be giving a 1d numpy array of tuples, but cdist needs a 2d numpy array. you can convert it to 2d array by using np.array([*currency.bc.values])

see the example below

from scipy.spatial import distance
import pandas as pd
import numpy as np

mypoint = [(0, 0)]
df = pd.DataFrame({'coord1': [(0,10), (10,0)]})
#option 1    
print(distance.cdist(mypoint , np.array([*df.coord1.values]), 'euclidean'))
#option2 
print(distance.cdist(mypoint , df.coord1.values.tolist(), 'euclidean'))

results in

[[10. 10.]]
[[10. 10.]]