CODEWITHSUNDEEP
pythonpandasdataframedictionary
Amir Hossein F
Amir Hossein F

Reputation: 420

put the whole list in one dataframe column

I'm trying to create a dataframe from a dictionary:

dict = {'foo': [1, 2, 3, 4],
        'bar': [5, 6, 7, 8]}

and I use the below command to create the dataframe:

df = pd.DataFrame.from_dict(dict, orient='index')

but the output is like:

df:
           0  1  2  3
     foo   1  2  3  4
     bar   4  5  6  7

but I expect the output to be like below with only one column:

df:
           'column_name'
     foo    [1, 2, 3, 4]
     bar    [4, 5, 6, 7]

Upvotes: 1

Views: 1248

Answers (5)

C.Nivs
C.Nivs

Reputation: 13106

If you're in python3.6+ you can lean on the ordered nature of dictionaries to create a different dictionary of just column_name being the key:

d = {'foo': [1, 2, 3, 4], 'bar': [5, 6, 7, 8]}

df = pd.DataFrame([{'column_name': v} for v in d.values()], index=d.keys())

df
      column_name
foo  [1, 2, 3, 4]
bar  [5, 6, 7, 8]

Otherwise, I would definitely default to @user3483203's answer

Upvotes: 1

user3483203
user3483203

Reputation: 51165

You are passing a dictionary containing "list-like" values. When passed to the DataFrame constructor, pandas interprets the keys of the dictionaries as Series labels, and the values in each list as new row values per Series.

When you use the from_dict class method, you are given the option of orientation, which allows you to specify whether or not the keys of the dictionary represent row or column labels, but the "list-like" values of the dictionary will still be interpreted either as new columns or new rows.

Using both of these methods also requires that the lengths of the values are uniform.

pd.DataFrame.from_dict(dct, orient='index')


     0  1  2  3
foo  1  2  3  4
bar  5  6  7  8


pd.DataFrame.from_dict(dct, orient='columns')


   foo  bar
0    1    5
1    2    6
2    3    7
3    4    8

Instead, you are interested in the 1-Dimensional pd.Series, which will take a dictionary and use each key as a row label, and each value as a row value.

pd.Series(dct)


foo    [1, 2, 3, 4]
bar    [5, 6, 7, 8]
dtype: object

Per my comment above, if you are interested in a DataFrame, you can use to_frame, which will maintain the values stored in the Series.

pd.Series(dct).to_frame('column_name')


      column_name
foo  [1, 2, 3, 4]
bar  [5, 6, 7, 8]

Upvotes: 2

Nev1111
Nev1111

Reputation: 1049

You can try:

frame=pd.Series(dict).rename('column_name')

Upvotes: 1

Andy L.
Andy L.

Reputation: 25259

Using the dict to create series and convert to dataframe:

dct = {'foo': [1, 2, 3, 4],
       'bar': [5, 6, 7, 8]}

pd.Series(dct).to_frame('column_name')

Out[937]:
      column_name
bar  [5, 6, 7, 8]
foo  [1, 2, 3, 4]

Note: please don't use dict as the variable name. It's a bad practice.

Upvotes: 1

René
René

Reputation: 4827

pd.DataFrame({'col_name': pd.Series(dict)})

Result:

         col_name
foo  [1, 2, 3, 4]
bar  [5, 6, 7, 8]

Upvotes: 1

Related Questions