to_pandas_with_time_index#

causalpy.input_data.to_pandas_with_time_index(data, time_column=None, *, argument_name='data')[source]#

Convert a dataframe-like input and put its time axis on the index.

Experiments that compare observations against a treatment_time need a time axis. Pandas callers have historically supplied it as the dataframe index. Dataframes from other libraries have no index, so those callers must name the column that holds the time axis.

Parameters:

  • data (NativeDataFrame) – Any eager dataframe supported by Narwhals.

  • time_column (str | None) – Column holding the time axis. When given, it becomes the index. When None, the pandas index of data is used, which requires a pandas input.

  • argument_name (str) – Name of the calling argument, used in error messages.

Returns:

A pandas dataframe indexed by the time axis.

Return type:

pandas.DataFrame

Raises:

DataException – If time_column is missing from the data, if a non-pandas input arrives without a time_column, if time_column is given for a dataframe that already carries a meaningful index, or if the resulting time axis has duplicates or is not sorted.

Examples

>>> import pandas as pd
>>> from causalpy.input_data import to_pandas_with_time_index
>>> frame = pd.DataFrame({"t": [1, 2], "y": [3, 4]})
>>> result = to_pandas_with_time_index(frame, time_column="t")
>>> result.index.name
't'
>>> result.index.tolist()
[1, 2]
>>> result.columns.tolist()
['y']