Pandas: make new column and return list of values from other column

Question

Suppose I have two columns

d = {'a_lower': [1, 2], 'a_upper': [3, 4]}
df = pd.DataFrame(data=d)

   a_lower  a_upper
0        1        3
1        2        4

I want to have 3rd column that returns a list from the values of the two columns

   a_lower  a_upper            a
0        1        3        [1, 3]
1        2        4        [2, 4]

I tried this

df['a'] = [df['a_lower'], df['a_upper']]

I got different result

    a_lower a_upper a
0   1   3   0 1 1 2 Name: a_lower, dtype: int64
1   2   4   0 3 1 4 Name: a_upper, dtype: int64

How to do if correcty? I am trying to return an array of dict oriented in ‘records’

[{'a_lower': 1,
  'a_upper': 3,
  'a': [1, 3]
 },
...
]

Asked By: Nikko

||

Source

Answer 1

Use list comprehension:

df['a'] = [[a,b] for a,b in zip(df['a_lower'], df['a_upper'])]

print (df)
   a_lower  a_upper       a
0        1        3  [1, 3]
1        2        4  [2, 4]

Or select only both columns and convert to numpy array by DataFrame.to_numpy and then to lists:

df['a'] = df[['a_lower','a_upper']].to_numpy().tolist()
#if only 2 columns DataFrame
#df['a'] = df.to_numpy().tolist()
print (df)
 
    a_lower  a_upper       a
0        1        3  [1, 3]
1        2        4  [2, 4]

Or use DataFrame.to_dict with split parameter and select data key:

df['a'] = df[['a_lower','a_upper']].to_dict('split')['data']

print (df)
 
    a_lower  a_upper       a
0        1        3  [1, 3]
1        2        4  [2, 4]

print (df[['a_lower','a_upper']].to_dict('split'))
{'index': [0, 1], 'columns': ['a_lower', 'a_upper'], 'data': [[1, 3], [2, 4]]}

Answered By: jezrael

Answer 2

Use numpy with to_numpy:

df['a'] = df.to_numpy().tolist()

or to limit to specific columns:

df['a'] = df[['a_lower', 'a_upper']].to_numpy().tolist()

Alternative, using a dictionary with to_dict:

df['a'] = df[['a_lower', 'a_upper']].to_dict('list').values()

Output:

   a_lower  a_upper       a
0        1        3  [1, 3]
1        2        4  [2, 4]

Answered By: mozway

Pandas: make new column and return list of values from other column

Question:

Answers: