列表列表转换为 pandas DataFrame

Question

I have a list of lists that looks like this:

[[('category', 'evaluation'), ('polarity', 'pos'), ('strength', '1'), ('type', 'good')],
[('category', 'intensifier'), ('type', 'shifter')],
[('category', 'evaluation'), ('polarity', 'pos'), ('strength', '2'), ('type', 'good')],

Note that not all lists contain all the attributes.

I would like, if possible, to convert this to a DataFrame, where each list represents a new row and the names of the columns would be given by the first element (eg 'category' , 'polarity' , 'strength' , 'type' ). In the end, the DataFrame should look like this:

           category     polarity     strength     type
df[0]:    evaluation      pos           1         good
df[1]:    intensifier     NaN          NaN       shifter
df[2]:    evaluation      pos           2         good

Any help would be greatly appreciated.

Answer 1

You could transform each list into a dictionary:

import pandas as pd

data = [[('category', 'evaluation'), ('polarity', 'pos'), ('strength', '1'), ('type', 'good')],
[('category', 'intensifier'), ('type', 'shifter')],
[('category', 'evaluation'), ('polarity', 'pos'), ('strength', '2'), ('type', 'good')]]

df = pd.DataFrame(data=[dict(e) for e in data])

print(df)

Output

      category polarity strength     type
0   evaluation      pos        1     good
1  intensifier      NaN      NaN  shifter
2   evaluation      pos        2     good