fix(pandas): use arrays for values and indices in time_features

[elephaint]

The following example fails without the fix:

import pandas as pd
from utilsforecast.feature_engineering import time_features

df_cv = pd.DataFrame({'DATE': {0: '2022-02-01',
  1: '2022-02-01',
  2: '2022-02-01',
  3: '2022-02-01',
  4: '2022-02-01'},
 'STORE_ID': {0: '51a91370-5e6e-11ea-bdc8-4f8d0b878ff3:Store',
  1: '8f4e06d0-6260-11eb-a46e-abb30e01a929:Store',
  2: '9c1da2e0-6020-11eb-8f38-59a0094a41b9:Store',
  3: 'c4df2260-60ef-11eb-a4b6-e18772606843:Store',
  4: '1433b940-9475-11ea-9faa-41f6926ddb7f:Store'},
 'NET_SALES': {0: 0.0, 1: 0.0, 2: 0.0, 3: 0.0, 4: 0.0}})

id_col = "STORE_ID"
time_col = "DATE"
target_col = "NET_SALES"

freq = "D"
h=28  # Short Term Foreasting
df_cv[time_col] = pd.to_datetime(df_cv[time_col])

df, X_df = time_features(
    df=df_cv.head(),
    freq=freq,
    features=["week", "dayofweek"],  # gives error
    id_col=id_col,
    time_col=time_col,
)

which leads to the error:

ValueError: cannot reindex on an axis with duplicate labels

which seems due to the weekly created feature not being an index.

Versions:

• utilsforecast 0.2.10
• pandas 2.2.3

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB