pandas.tseries.offsets.DateOffset.rollback#

DateOffset.rollback(dt)#

Roll provided date backward to next offset only if not on offset.

If the provided date is already on an offset, it is returned unchanged. Otherwise, the date is rolled backward to the previous offset boundary.

Parameters:

Returns:

Timestamp: Rolled timestamp if not on offset, otherwise unchanged timestamp.

See also

rollforward: Roll provided date forward to next offset only if not on offset.
is_on_offset: Return boolean whether a timestamp intersects with this frequency.

Examples

>>> ts = pd.Timestamp("2025-01-15 09:00:00")
>>> offset = pd.tseries.offsets.MonthEnd()

Timestamp is not on the offset (not a month end), so it rolls backward:

>>> offset.rollback(ts)
Timestamp('2024-12-31 00:00:00')

If the timestamp is already on the offset, it remains unchanged:

>>> ts_on_offset = pd.Timestamp("2025-01-31")
>>> offset.rollback(ts_on_offset)
Timestamp('2025-01-31 00:00:00')