Is it possible to assign a MultiIndex to an empty DataFrame?

Question

I've created a MultiIndex and want to create an empty data frame with the MultiIndex as the column labels and then fill the matrix later, but it just returns a regular index with tuples. This code

import pandas as pd

sizes = ['0_200', '200_700', '700_1400', '1400_2300', '2300_4000']
years = [2016, 2020, 2025, 2030, 2035, 2040, 2045, 2050]

multidx = pd.MultiIndex.from_product([sizes, years], names=['size', 'year'])
print(multidx)
df = pd.DataFrame(columns=multidx)
print(df)

yields

MultiIndex(levels=[['0_200', '1400_2300', '200_700', '2300_4000', '700_1400'], [2016, 2020, 2025, 2030, 2035, 2040, 2045, 2050]],
           labels=[[0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 4, 4, 4, 4, 4, 4, 4, 4, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3], [0, 1, 2, 3, 4, 5, 6, 7, 0, 1, 2, 3, 4, 5, 6, 7, 0, 1, 2, 3, 4, 5, 6, 7, 0, 1, 2, 3, 4, 5, 6, 7, 0, 1, 2, 3, 4, 5, 6, 7]],
           names=['size', 'year'])

Empty DataFrame
Columns: [(0_200, 2016), (0_200, 2020), (0_200, 2025), (0_200, 2030), (0_200, 2035), (0_200, 2040), (0_200, 2045), (0_200, 2050), (200_700, 2016), (200_700, 2020), (200_700, 2025), (200_700, 2030), (200_700, 2035), (200_700, 2040), (200_700, 2045), (200_700, 2050), (700_1400, 2016), (700_1400, 2020), (700_1400, 2025), (700_1400, 2030), (700_1400, 2035), (700_1400, 2040), (700_1400, 2045), (700_1400, 2050), (1400_2300, 2016), (1400_2300, 2020), (1400_2300, 2025), (1400_2300, 2030), (1400_2300, 2035), (1400_2300, 2040), (1400_2300, 2045), (1400_2300, 2050), (2300_4000, 2016), (2300_4000, 2020), (2300_4000, 2025), (2300_4000, 2030), (2300_4000, 2035), (2300_4000, 2040), (2300_4000, 2045), (2300_4000, 2050)]
Index: []
[0 rows x 40 columns]

I'm using Spyder 3.1.4, Python 3.6.0 64bits on Windows. Pandas version 0.19.2.

It is multi index if you print df.columns. Or try df.loc[0,:] = np.arange(40) and then print(df). It's likely that's just how an empty data frame is printed. — akuiper
– akuiper, Commented Jun 6, 2017 at 14:46
Yep, it looks like you were successful in your attempts to create an empty df with multiindex in columns. — Scott Boston
– Scott Boston, Commented Jun 6, 2017 at 15:12

Ringvold · Accepted Answer · 2017-06-07 12:28:47Z

1

Thank you Psidom and Scott Boston. You're right that the multiindex is there, it just prints as tuples before the data frame is filled. I'll leave this question here because it would have been really handy for me about 24 hours ago. (I'm new here though, so if this counts as a non-question and I should rather delete it, let me know!)

answered Jun 7, 2017 at 12:28

Ringvold

637 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Is it possible to assign a MultiIndex to an empty DataFrame?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related