list of dicts with nested dicts to pandas dataframe

Question

I have a list of dicts. Some of the objects in these dicts are dicts themselves. I would like to turn this into a dataframe. This is what I currently get.

In[163]:
res = [{'A':1,'B':{'C':3,'D':4}},{'A':5,'B':{'C':7,'D':8}}]
pd.DataFrame(res)

Out[163]:
    A   B
0   1   {u'C': 3, u'D': 4}
1   5   {u'C': 7, u'D': 8}

This is what I want:

    A   BC  BD
0   1   3   4
1   5   7   8

I don't really care about the column header names. They can be whatever.

Answer 1

import pandas as pd
res = [{'A':1,'B':{'C':3,'D':4}},{'A':5,'B':{'C':7,'D':8}}]
df = pd.DataFrame(res)

def col2frame(col):
    try:
        res = pd.DataFrame.from_records(col)
        res.columns = [col.name+name for name in res.columns]
        return res
    except TypeError:
        return col

df = pd.concat([col2frame(df[col]) for col in df], axis=1)
print(df)

yields

   A  BC  BD
0  1   3   4
1  5   7   8