CODEWITHSUNDEEP
pythonpandasdictionarydata-manipulation
baxx
baxx

Reputation: 4695

Create a pandas dataframe from a dictionary of dictionaries

I have a dictionary of dictionaries similar to the following:

data_dict = {
    'first_entry': {'a': 345, 'b': 8484}, 
    'second_entry': {'a': 3423, 'b': 848}
}

I would like to create a dataframe of all these values, as follows:

pd.DataFrame(
    [list(data_dict[key].values()) for key in data_dict.keys()],
    columns = list(data_dict[list(data_dict.keys())[0]].keys()))

I'm a bit concerned about the approach taken here with respect to accessing the keys and such.

Note - in the above the values first_entry and second_entry are not reliable, but the values of a and b are reliable. In the actual data I have ~500 or so nested dictionaries, (so first_entry ... five_hundredth_entry using the above syntax).

Upvotes: 1

Views: 111

Answers (2)

BENY
BENY

Reputation: 323226

You can simply pass the nested dictionary to the DataFrame constructor then transpose and drop the index:

pd.DataFrame(d).T.reset_index(drop=True)
      a     b
0   345  8484
1  3423   848

Upvotes: 2

ansev
ansev

Reputation: 30920

You only need DataFrame.from_dict with orient='index'. We can reset the index at the end optionally.

new_df = pd.DataFrame.from_dict(data_dict, orient='index').reset_index(drop=True)
print(new_df)

Output

      a     b
0   345  8484
1  3423   848

Upvotes: 3

Related Questions