bulum.io.csv_io module

Functions for reading CSVs, particularly time-series CSVs.

read_ts_csv(filename: str | PathLike, date_format=None, df=None, colprefix=None, allow_nonnumeric=False, assert_date=True, **kwargs) → TimeseriesDataframe

Reads a daily timeseries csv into a DataFrame, and sets the index to string dates in the “%Y-%m-%d” format. The method assumes the first column are dates.

Parameters:

  • filename (str or PathLike) – Path to the CSV file to read.

  • date_format (str, optional) – defaults to “%d/%m/%Y” as per Fors. Other common formats include “%Y-%m-%d”, “%Y/%m/%d”.

  • df (pd.DataFrame, optional) – If provided, the reader will append columns to this dataframe. Defaults to None.

  • colprefix (str, optional) – If provided, the reader will prepend this to column name. Defaults to None.

  • allow_nonnumeric (bool, optional) – If False, raises if any column is not numeric. Default is False.

  • assert_date (bool, optional) – If True, asserts the date index conforms to "%Y-%m-%d". Default is True.

Return type:

utils.TimeseriesDataframe

Raises:

  • TypeError – If a column is not numeric and allow_nonnumeric is False.

  • ValueError – If the new date range does not overlap with df.

write_ts_csv(df: DataFrame, filename: str, *args, **kwargs) → None

Write a dataframe to CSV via pandas.DataFrame.to_csv().

Parameters: