[FIXED] Pandas dataframe from unique numpy elements

Issue

I want to construct a pandas dataframe where columns have elements from specified arrays with unique elements only. I want to find the most efficient pythonic way of doing this.

For example, I have the following numpy arrays as input:

a = np.array(["a1"])
b = np.array(["b1", "b2"])
c = np.array(["c1", "c2", "c3"])

Below is how I want my Pandas dataframe to be like

    a   b   c
0  a1  b1  c1
1  a1  b1  c2
2  a1  b1  c3
3  a1  b2  c1
4  a1  b2  c2
5  a1  b2  c3

Below is the code that I am using:

import pandas as pd

hashmap = {"a":[], "b":[], "c":[]}

for a_elem in a:
    for b_elem in b:
        for c_elem in c:
            hashmap["a"] += [a_elem]
            hashmap["b"] += [b_elem]
            hashmap["c"] += [c_elem]

df = pd.DataFrame.from_dict(hashmap)

How can I make this code more efficient?

Solution

Use itertools.product:

from itertools import product

df = pd.DataFrame(product(a, b, c), columns=['a', 'b', 'c'])

Output:

    a   b   c
0  a1  b1  c1
1  a1  b1  c2
2  a1  b1  c3
3  a1  b2  c1
4  a1  b2  c2
5  a1  b2  c3

Alternative with MultiIndex.from_product:

df = pd.MultiIndex.from_product([a, b, c], names=['a', 'b', 'c']
                                ).to_frame(index=False)

Answered By - mozway