pd_to_tensor

# Import statement

from hottbox.pdtools import pd_to_tensor

pd_to_tensor(df, keep_index=True)

Represent multi-index pandas dataframe as a tensor

Parameters

dfpd.DataFrame

Multi-index dataframe with only one column of data

keep_indexbool

Keep level values of dataframe multi-index

Returns

tensorTensor

Examples

>>> import numpy as np
>>> import pandas as pd
>>> from hottbox.pdtools import pd_to_tensor
>>> data = {'Year': [2005, 2005, 2005, 2005, 2010, 2010, 2010, 2010],
...         'Month': ['Jan', 'Jan', 'Feb', 'Feb', 'Jan', 'Jan', 'Feb', 'Feb'],
...         'Day': ['Mon', 'Wed', 'Mon', 'Wed', 'Mon', 'Wed', 'Mon', 'Wed'],
...         'Population': np.arange(8)
...         }
>>> df = pd.DataFrame.from_dict(data)
>>> df.set_index(["Year", "Month", "Day"], inplace=True)
>>> print(df)
                    Population
    Year Month Day
    2005 Jan   Mon           0
               Wed           1
         Feb   Mon           2
               Wed           3
    2010 Jan   Mon           4
               Wed           5
         Feb   Mon           6
               Wed           7
>>> tensor = pd_to_tensor(df)
>>> print(tensor.data)
    [[[0 1]
      [2 3]]
     [[4 5]
      [6 7]]]
>>> print(tensor)
    This tensor is of order 3 and consists of 8 elements.
    Sizes and names of its modes are (2, 2, 2) and ['Year', 'Month', 'Day'] respectively.
>>> tensor.modes
    [Mode(name='Year', index=[2005, 2010]),
     Mode(name='Month', index=['Jan', 'Feb']),
     Mode(name='Day', index=['Mon', 'Wed'])]
>>> tensor = pd_to_tensor(df, keep_index=False)
>>> tensor.modes
    [Mode(name='Year', index=None),
     Mode(name='Month', index=None),
     Mode(name='Day', index=None)]