Can dataframe’s cells hold nested dictionary as value?

user_8294738 0
Viewed 67

dict:

meta_data = {
                "attr1":"meta_value1", 
                "attr2": {
                             "key1": "value1",
                             "key2": "value2",
                             "key3": "value3",
                         },
             }

then I want to create a dataframe by doing this:

df = pd.DataFrame({"df_key1": "df_value1", "df_key2": "df_value2", 
                   "df_key3": "df_value3", "meta_data": meta_data}, index[0])

I got:

        df_key1        df_key2        df_key3        meta_data
0       df_value1      df_value2      df_value3      NaN

columns 'meta_data' valued NaN.

I also tried

df_dict = {"df_key1": "df_value1", "df_key2": "df_value2", 
           "df_key3": "df_value3", "meta_data": meta_data}

df = pd.DataFrame.from_dict(df_dict)

I got:

            df_key1        df_key2        df_key3        meta_data
attr1       df_value1      df_value2      df_value3      meta_value1
attr2       df_value1      df_value2      df_value3      {'key1': 'value1', 'key2': 'value2', 'key3': 'value3'}

what I really want is:

        df_key1        df_key2        df_key3        meta_data
0       df_value1      df_value2      df_value3      {'attr1': 'meta_value1', 'attr2': {'key1': 'value1', 'key2': 'value2', 'key3': 'value3'}}

a nested dictionary as the value in dataframe's cell,

what should I do?

All answers can solve my problem, I picked the most agreeable one for myself, thank you all.

If you have any suggestions from any reasonable perspectives like code style or else welcome to comment.

pandas dataframe numpy python
3 Answers

use pd.Series then pd.DataFrame


df = pd.Series({"df_key1": "df_value1", "df_key2": "df_value2", 
                  "df_key3": "df_value3", "meta_data": meta_data})
df.to_frame().T
user_60063830

You should try creating dataframe with key value pairs using .items() from dictionary.

Setup:

import pandas as pd

meta_data = {
                "attr1":"meta_value1", 
                "attr2": {
                             "key1": "value1",
                             "key2": "value2",
                             "key3": "value3",
                         },
             }


df_dict = {"df_key1": "df_value1", "df_key2": "df_value2", 
           "df_key3": "df_value3", "meta_data": meta_data}

And this will give the required dataframe:

pd.DataFrame({k: [v] for k, v in df_dict.items()})
user_148844540

Upon changing the DataFrame creation , wrapping meta_data inside a list solves the required output -

>>> df_dict = {"df_key1": "df_value1", "df_key2": "df_value2","df_key3": "df_value3", "meta_data": [meta_data]}
>>> df = pd.DataFrame.from_dict(df_dict)
>>> df
     df_key1    df_key2    df_key3                                          meta_data
0  df_value1  df_value2  df_value3  {'attr1': 'meta_value1', 'attr2': {'key1': 'va...
>>> df.meta_data.values
array([{'attr1': 'meta_value1', 'attr2': {'key1': 'value1', 'key2': 'value2', 'key3': 'value3'}}],
      dtype=object)
user_91089120
Related