Create a pandas dataframe from a nested lists of unequal lengths

user_5236124 0
Viewed 6545

So I have a list as follows:

aa = ['aa1', 'aa2', 'aa3', 'aa4', 'aa5']
bb = ['bb1', 'bb2', 'bb3', 'bb4']
cc = ['cc1', 'cc2', 'cc3']

Which is then created into a nested list: 

nest = [aa, bb, cc]

I want to create a dataframe as follows:

aa   bb   cc
aa1  bb1  cc1
aa2  bb2  cc2
aa3  bb3  cc3
aa4  bb4  nan
aa5  nan  nan

I've tried:

pd.DataFrame(nest, columns=['aa', 'bb', cc'])

But results is such that, each list is being written as a row (as opposed to a column)

pandas list nested python
3 Answers

The zip_longest function from itertools does this:

>>> import itertools, pandas
>>> pandas.DataFrame((_ for _ in itertools.zip_longest(*nest)), columns=['aa', 'bb', 'cc'])
    aa    bb    cc
0  aa1   bb1   cc1
1  aa2   bb2   cc2
2  aa3   bb3   cc3
3  aa4   bb4  None
4  aa5  None  None

If you have an older version of pandas you may need to wrap zip_longest in a list constructor. On older Python you may need to call izip_longest instead of zip_longest.

user_4440360

Option 1 

pd.DataFrame(nest, ['aa', 'bb', 'cc']).T

    aa    bb    cc
0  aa1   bb1   cc1
1  aa2   bb2   cc2
2  aa3   bb3   cc3
3  aa4   bb4  None
4  aa5  None  None

Option 2
Homebrew zip_longest 

f = lambda x, n: x[n] if n < len(x) else None
n, m = max(map(len, nest)), len(nest)

pd.DataFrame(
    [[f(j, i) for j in nest] for i in range(n)],
    columns=['aa', 'bb', 'cc']
)

    aa    bb    cc
0  aa1   bb1   cc1
1  aa2   bb2   cc2
2  aa3   bb3   cc3
3  aa4   bb4  None
4  aa5  None  None
user_23366540

Or maybe 

pd.DataFrame(data={'value':nest},index=['aa', 'bb', 'cc']).value.apply(pd.Series).T
Out[1297]: 
    aa   bb   cc
0  aa1  bb1  cc1
1  aa2  bb2  cc2
2  aa3  bb3  cc3
3  aa4  bb4  NaN
4  aa5  NaN  NaN
user_79645270
Related