pandas examples

Calendar accepts pandas date-like inputs directly. The important point is that the top-level API remains NumPy-backed: pandas values are accepted as inputs, while outputs remain NumPy scalars or NumPy arrays: dates are datetime64[D] with NaT for missing values, counts are float64 with NaN for missing values, and isbizday vectors are masked arrays with fill_value=False.

import pandas as pd

from bizdays import Calendar

cal = Calendar.load(name="ANBIMA")

Pandas DatetimeIndex values work directly with isbizday.

dates = pd.to_datetime(["2014-01-12", "2014-01-13", None])
cal.isbizday(dates)

masked_array(data=[False, True, --],
             mask=[False, False,  True],
       fill_value=False)

Scalar pandas timestamps are accepted and still produce NumPy scalars.

result = cal.following(pd.Timestamp("2015-12-25"))
result, type(result)

(np.datetime64('2015-12-28'), numpy.datetime64)

Vectorized business-day calculations fit naturally into a DataFrame workflow.

df = pd.DataFrame(
    {
        "trade_date": pd.to_datetime(["2024-01-02", "2024-01-03"]),
        "settlement_date": pd.to_datetime(["2024-01-05", "2024-01-10"]),
    }
)
df["bdays"] = cal.bizdays(df["trade_date"], df["settlement_date"])
df

	trade_date	settlement_date	bdays
0	2024-01-02	2024-01-05	3.0
1	2024-01-03	2024-01-10	5.0

Results are positional NumPy arrays: the Series index is not kept. Assign a result straight to a column, as above. When you need a Series, pass index=.

df2 = df.set_index(pd.Index(["a", "b"]))
pd.Series(cal.bizdays(df2["trade_date"], df2["settlement_date"]), index=df2.index)

a    3.0
b    5.0
dtype: float64

Dates outside the calendar range raise DateOutOfRange. Check input dates first with inrange, and compute only the rows inside the range. filled(False) turns missing flags into False.

loans = pd.DataFrame({"start": ["2024-01-02", "2024-01-03"], "maturity": ["2024-06-15", "2100-01-03"]})
ok = cal.inrange(loans.maturity).filled(False)
loans.loc[ok, "du"] = cal.bizdays(loans.start[ok], loans.maturity[ok])
loans

	start	maturity	du
0	2024-01-02	2024-06-15	113.0
1	2024-01-03	2100-01-03	NaN

getdate also works with vectorized month references that come from pandas objects.

months = pd.Index(["2024-01", "2024-02", "2024-03"])
cal.getdate("15th bizday", months)

array(['2024-01-22', '2024-02-23', '2024-03-21'], dtype='datetime64[D]')

Missing pandas inputs are preserved as NaT (dates), NaN (counts) or masked values (isbizday), depending on the method.

cal.bizdays(pd.to_datetime(["2024-01-02", None]), pd.Timestamp("2024-01-05"))

array([ 3., nan])