[FIXED] Count values from different columns of a dataframe

Issue

Let's say I have the following dataframe.

import pandas as pd

data = {
  'home': ['team1', 'team2', 'team3', 'team2'],
  'away': ['team2', 'team3', 'team1', 'team1']
 }
df = pd.DataFrame(data)

How can I count the number of time each element (team) appears in both columns ? The expected result is

team1    3
team2    3
team3    2

Solution

You can concatenate the columns and use .value_counts method:

out = pd.concat([df['home'], df['away']]).value_counts()

Output:

team1    3
team2    3
team3    2
dtype: int64

You can also get the underlying numpy array, flatten it, find unique values and their counts, wrap it in a dictionary (this is by far the fastest method):

out = dict(np.array(np.unique(df.values.flatten(), return_counts=True)).T)

Output:

{'team1': 3, 'team2': 3, 'team3': 2}

Answered By - enke