Summing Numpy Arrays

Question

I have one numpy array of the form:

np.Array1 = 

[
   ['2019-12-01' '0.03555' '0.03' '0.03' '0.03'],
   ['2019-12-02' '0.03' '0.03' '1' '0.03']
]

and a second:

np.Array2 = 

[
   array(['2019-12-01', '1', '1', '1', '1']),
   array(['2019-12-02', '1', '1', '1', '20'])
]

Is there a way for me so do something along the lines of:

Sum each element where npArray1.col1 = npArray2.col1 - i.e : when the dates are the same add element by element (excluding the date)

'2019-12-01' = '2019-12-01' so [0.03555+1, 0.03+1, 0.03+1, 0.03+1]

I get that I am probs going about this the wrong way, by varying the types in the same array

Any advice is appreciated on a better way of add values based on conditional logic.

sacuL · Accepted Answer · 2020-02-12 13:53:32Z

2

You could do this by converting your arrays to pandas dataframes with the date as the index, and using add:

import numpy as np
import pandas as pd

a1 = np.array( [['2019-12-01', '0.03555', '0.03', '0.03', '0.03'],
                ['2019-12-02', '0.03', '0.03', '1', '0.03']] )

a2 = np.array([['2019-12-01', '1', '1', '1', '1'],
               ['2019-12-02', '1', '1', '1', '20']])

# convert to dataframe and set the date as the index
# also convert to floats:
df1 = pd.DataFrame(a1).set_index(0).astype(float)
df2 = pd.DataFrame(a2).set_index(0).astype(float)


df1.add(df2, fill_value = 0)

You can then get it back as a numpy array in your original format by converting back to string, reseting the index, and taking the values:

df1.add(df2, fill_value = 0).astype(str).reset_index().values

answered Feb 12, 2020 at 13:53

sacuL

51.6k9 gold badges88 silver badges115 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Bob Over a year ago

Thanks for this, I did something similar, I converted the original source of NumPy arrays to return an array of dicts instead, then pd.DataFrame(self.buildArrayOfDicts()).explode('dates') - I then did a df.set_index('dates) making it easier to sum() later down the line

kalzso · Accepted Answer · 2020-02-12 14:05:03Z

This worked for me, and it not requires pandas:

import numpy as np

a1 = [
    np.array(['2019-12-01', '0.03555', '0.03', '0.03', '0.03']),
    np.array(['2019-12-02', '0.03', '0.03', '1', '0.03'])
]

a2 = [
    np.array(['2019-12-01', '1', '1', '1', '1']),
    np.array(['2019-12-02', '1', '1', '1', '20'])
]

def array_adder(A, B):
    C = []
    for outer, outer2 in zip(A, B):
        temp = []
        if len(outer) == len(outer2):
            for inner, inner2 in zip(outer, outer2):
                if len(temp) < 1:
                    temp.append(inner)
                else:
                    temp.append(float(inner) + float(inner2))
            C.append(temp)
        else:
            print('Arrays must have the same length..')
            break
    return C

print(array_adder(a1, a2))

Collectives™ on Stack Overflow

Summing Numpy Arrays

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related