ENH: Timestamp.replace support non-nano (pandas-dev#47312)

Author: jbrockmendel

pandas/_libs/tslibs/conversion.pxd

@@ -27,7 +27,8 @@ cdef _TSObject convert_to_tsobject(object ts, tzinfo tz, str unit,
                                    int32_t nanos=*)
 
 cdef _TSObject convert_datetime_to_tsobject(datetime ts, tzinfo tz,
-                                            int32_t nanos=*)
+                                            int32_t nanos=*,
+                                            NPY_DATETIMEUNIT reso=*)
 
 cdef int64_t get_datetime64_nanos(object val) except? -1
 

pandas/_libs/tslibs/conversion.pyx

@@ -31,6 +31,7 @@ from cpython.datetime cimport (
 import_datetime()
 
 from pandas._libs.tslibs.base cimport ABCTimestamp
+from pandas._libs.tslibs.dtypes cimport periods_per_second
 from pandas._libs.tslibs.np_datetime cimport (
     NPY_DATETIMEUNIT,
     NPY_FR_ns,
@@ -40,11 +41,14 @@ from pandas._libs.tslibs.np_datetime cimport (
     dtstruct_to_dt64,
     get_datetime64_unit,
     get_datetime64_value,
+    get_implementation_bounds,
     get_unit_from_dtype,
     npy_datetime,
     npy_datetimestruct,
+    npy_datetimestruct_to_datetime,
     pandas_datetime_to_datetimestruct,
     pydatetime_to_dt64,
+    pydatetime_to_dtstruct,
     string_to_dts,
 )
 
@@ -307,11 +311,15 @@ cdef maybe_localize_tso(_TSObject obj, tzinfo tz, NPY_DATETIMEUNIT reso):
     if obj.value != NPY_NAT:
         # check_overflows needs to run after _localize_tso
         check_dts_bounds(&obj.dts, reso)
-        check_overflows(obj)
+        check_overflows(obj, reso)
 
 
-cdef _TSObject convert_datetime_to_tsobject(datetime ts, tzinfo tz,
-                                            int32_t nanos=0):
+cdef _TSObject convert_datetime_to_tsobject(
+    datetime ts,
+    tzinfo tz,
+    int32_t nanos=0,
+    NPY_DATETIMEUNIT reso=NPY_FR_ns,
+):
     """
     Convert a datetime (or Timestamp) input `ts`, along with optional timezone
     object `tz` to a _TSObject.
@@ -327,13 +335,15 @@ cdef _TSObject convert_datetime_to_tsobject(datetime ts, tzinfo tz,
         timezone for the timezone-aware output
     nanos : int32_t, default is 0
         nanoseconds supplement the precision of the datetime input ts
+    reso : NPY_DATETIMEUNIT, default NPY_FR_ns
 
     Returns
     -------
     obj : _TSObject
     """
     cdef:
         _TSObject obj = _TSObject()
+        int64_t pps
 
     obj.fold = ts.fold
     if tz is not None:
@@ -342,34 +352,35 @@ cdef _TSObject convert_datetime_to_tsobject(datetime ts, tzinfo tz,
         if ts.tzinfo is not None:
             # Convert the current timezone to the passed timezone
             ts = ts.astimezone(tz)
-            obj.value = pydatetime_to_dt64(ts, &obj.dts)
+            pydatetime_to_dtstruct(ts, &obj.dts)
             obj.tzinfo = ts.tzinfo
         elif not is_utc(tz):
             ts = _localize_pydatetime(ts, tz)
-            obj.value = pydatetime_to_dt64(ts, &obj.dts)
+            pydatetime_to_dtstruct(ts, &obj.dts)
             obj.tzinfo = ts.tzinfo
         else:
             # UTC
-            obj.value = pydatetime_to_dt64(ts, &obj.dts)
+            pydatetime_to_dtstruct(ts, &obj.dts)
             obj.tzinfo = tz
     else:
-        obj.value = pydatetime_to_dt64(ts, &obj.dts)
+        pydatetime_to_dtstruct(ts, &obj.dts)
         obj.tzinfo = ts.tzinfo
 
-    if obj.tzinfo is not None and not is_utc(obj.tzinfo):
-        offset = get_utcoffset(obj.tzinfo, ts)
-        obj.value -= int(offset.total_seconds() * 1e9)
-
     if isinstance(ts, ABCTimestamp):
-        obj.value += <int64_t>ts.nanosecond
         obj.dts.ps = ts.nanosecond * 1000
 
     if nanos:
-        obj.value += nanos
         obj.dts.ps = nanos * 1000
 
-    check_dts_bounds(&obj.dts)
-    check_overflows(obj)
+    obj.value = npy_datetimestruct_to_datetime(reso, &obj.dts)
+
+    if obj.tzinfo is not None and not is_utc(obj.tzinfo):
+        offset = get_utcoffset(obj.tzinfo, ts)
+        pps = periods_per_second(reso)
+        obj.value -= int(offset.total_seconds() * pps)
+
+    check_dts_bounds(&obj.dts, reso)
+    check_overflows(obj, reso)
     return obj
 
 
@@ -401,7 +412,7 @@ cdef _TSObject _create_tsobject_tz_using_offset(npy_datetimestruct dts,
     obj.tzinfo = pytz.FixedOffset(tzoffset)
     obj.value = tz_localize_to_utc_single(value, obj.tzinfo)
     if tz is None:
-        check_overflows(obj)
+        check_overflows(obj, NPY_FR_ns)
         return obj
 
     cdef:
@@ -515,13 +526,14 @@ cdef _TSObject _convert_str_to_tsobject(object ts, tzinfo tz, str unit,
     return convert_datetime_to_tsobject(dt, tz)
 
 
-cdef inline check_overflows(_TSObject obj):
+cdef inline check_overflows(_TSObject obj, NPY_DATETIMEUNIT reso=NPY_FR_ns):
     """
     Check that we haven't silently overflowed in timezone conversion
 
     Parameters
     ----------
     obj : _TSObject
+    reso : NPY_DATETIMEUNIT, default NPY_FR_ns
 
     Returns
     -------
@@ -532,15 +544,20 @@ cdef inline check_overflows(_TSObject obj):
     OutOfBoundsDatetime
     """
     # GH#12677
-    if obj.dts.year == 1677:
+    cdef:
+        npy_datetimestruct lb, ub
+
+    get_implementation_bounds(reso, &lb, &ub)
+
+    if obj.dts.year == lb.year:
         if not (obj.value < 0):
             from pandas._libs.tslibs.timestamps import Timestamp
             fmt = (f"{obj.dts.year}-{obj.dts.month:02d}-{obj.dts.day:02d} "
                    f"{obj.dts.hour:02d}:{obj.dts.min:02d}:{obj.dts.sec:02d}")
             raise OutOfBoundsDatetime(
                 f"Converting {fmt} underflows past {Timestamp.min}"
             )
-    elif obj.dts.year == 2262:
+    elif obj.dts.year == ub.year:
         if not (obj.value > 0):
             from pandas._libs.tslibs.timestamps import Timestamp
             fmt = (f"{obj.dts.year}-{obj.dts.month:02d}-{obj.dts.day:02d} "

pandas/_libs/tslibs/np_datetime.pxd

@@ -79,6 +79,7 @@ cdef int64_t dtstruct_to_dt64(npy_datetimestruct* dts) nogil
 cdef void dt64_to_dtstruct(int64_t dt64, npy_datetimestruct* out) nogil
 
 cdef int64_t pydatetime_to_dt64(datetime val, npy_datetimestruct *dts)
+cdef void pydatetime_to_dtstruct(datetime dt, npy_datetimestruct *dts)
 cdef int64_t pydate_to_dt64(date val, npy_datetimestruct *dts)
 cdef void pydate_to_dtstruct(date val, npy_datetimestruct *dts)
 
@@ -104,3 +105,6 @@ cpdef cnp.ndarray astype_overflowsafe(
 )
 
 cdef bint cmp_dtstructs(npy_datetimestruct* left, npy_datetimestruct* right, int op)
+cdef get_implementation_bounds(
+    NPY_DATETIMEUNIT reso, npy_datetimestruct *lower, npy_datetimestruct *upper
+)

pandas/_libs/tslibs/np_datetime.pyx

@@ -167,30 +167,34 @@ class OutOfBoundsTimedelta(ValueError):
     pass
 
 
+cdef get_implementation_bounds(NPY_DATETIMEUNIT reso, npy_datetimestruct *lower, npy_datetimestruct *upper):
+    if reso == NPY_FR_ns:
+        upper[0] = _NS_MAX_DTS
+        lower[0] = _NS_MIN_DTS
+    elif reso == NPY_FR_us:
+        upper[0] = _US_MAX_DTS
+        lower[0] = _US_MIN_DTS
+    elif reso == NPY_FR_ms:
+        upper[0] = _MS_MAX_DTS
+        lower[0] = _MS_MIN_DTS
+    elif reso == NPY_FR_s:
+        upper[0] = _S_MAX_DTS
+        lower[0] = _S_MIN_DTS
+    elif reso == NPY_FR_m:
+        upper[0] = _M_MAX_DTS
+        lower[0] = _M_MIN_DTS
+    else:
+        raise NotImplementedError(reso)
+
+
 cdef check_dts_bounds(npy_datetimestruct *dts, NPY_DATETIMEUNIT unit=NPY_FR_ns):
     """Raises OutOfBoundsDatetime if the given date is outside the range that
     can be represented by nanosecond-resolution 64-bit integers."""
     cdef:
         bint error = False
         npy_datetimestruct cmp_upper, cmp_lower
 
-    if unit == NPY_FR_ns:
-        cmp_upper = _NS_MAX_DTS
-        cmp_lower = _NS_MIN_DTS
-    elif unit == NPY_FR_us:
-        cmp_upper = _US_MAX_DTS
-        cmp_lower = _US_MIN_DTS
-    elif unit == NPY_FR_ms:
-        cmp_upper = _MS_MAX_DTS
-        cmp_lower = _MS_MIN_DTS
-    elif unit == NPY_FR_s:
-        cmp_upper = _S_MAX_DTS
-        cmp_lower = _S_MIN_DTS
-    elif unit == NPY_FR_m:
-        cmp_upper = _M_MAX_DTS
-        cmp_lower = _M_MIN_DTS
-    else:
-        raise NotImplementedError(unit)
+    get_implementation_bounds(unit, &cmp_lower, &cmp_upper)
 
     if cmp_npy_datetimestruct(dts, &cmp_lower) == -1:
         error = True
@@ -229,19 +233,23 @@ def py_td64_to_tdstruct(int64_t td64, NPY_DATETIMEUNIT unit):
     return tds  # <- returned as a dict to python
 
 
+cdef inline void pydatetime_to_dtstruct(datetime dt, npy_datetimestruct *dts):
+    dts.year = PyDateTime_GET_YEAR(dt)
+    dts.month = PyDateTime_GET_MONTH(dt)
+    dts.day = PyDateTime_GET_DAY(dt)
+    dts.hour = PyDateTime_DATE_GET_HOUR(dt)
+    dts.min = PyDateTime_DATE_GET_MINUTE(dt)
+    dts.sec = PyDateTime_DATE_GET_SECOND(dt)
+    dts.us = PyDateTime_DATE_GET_MICROSECOND(dt)
+    dts.ps = dts.as = 0
+
+
 cdef inline int64_t pydatetime_to_dt64(datetime val,
                                        npy_datetimestruct *dts):
     """
     Note we are assuming that the datetime object is timezone-naive.
     """
-    dts.year = PyDateTime_GET_YEAR(val)
-    dts.month = PyDateTime_GET_MONTH(val)
-    dts.day = PyDateTime_GET_DAY(val)
-    dts.hour = PyDateTime_DATE_GET_HOUR(val)
-    dts.min = PyDateTime_DATE_GET_MINUTE(val)
-    dts.sec = PyDateTime_DATE_GET_SECOND(val)
-    dts.us = PyDateTime_DATE_GET_MICROSECOND(val)
-    dts.ps = dts.as = 0
+    pydatetime_to_dtstruct(val, dts)
     return dtstruct_to_dt64(dts)
 
 

pandas/_libs/tslibs/timestamps.pyx

@@ -2203,9 +2203,6 @@ default 'raise'
             datetime ts_input
             tzinfo_type tzobj
 
-        if self._reso != NPY_FR_ns:
-            raise NotImplementedError(self._reso)
-
         # set to naive if needed
         tzobj = self.tzinfo
         value = self.value
@@ -2218,7 +2215,7 @@ default 'raise'
             value = tz_convert_from_utc_single(value, tzobj, reso=self._reso)
 
         # setup components
-        dt64_to_dtstruct(value, &dts)
+        pandas_datetime_to_datetimestruct(value, self._reso, &dts)
         dts.ps = self.nanosecond * 1000
 
         # replace
@@ -2265,12 +2262,16 @@ default 'raise'
                       'fold': fold}
             ts_input = datetime(**kwargs)
 
-        ts = convert_datetime_to_tsobject(ts_input, tzobj)
+        ts = convert_datetime_to_tsobject(ts_input, tzobj, nanos=0, reso=self._reso)
+        # TODO: passing nanos=dts.ps // 1000 causes a RecursionError in
+        #  TestTimestampConstructors.test_constructor; not clear why
         value = ts.value + (dts.ps // 1000)
         if value != NPY_NAT:
-            check_dts_bounds(&dts)
+            check_dts_bounds(&dts, self._reso)
 
-        return create_timestamp_from_ts(value, dts, tzobj, self._freq, fold)
+        return create_timestamp_from_ts(
+            value, dts, tzobj, self._freq, fold, reso=self._reso
+        )
 
     def to_julian_date(self) -> np.float64:
         """

pandas/tests/scalar/timestamp/test_unary_ops.py

@@ -19,6 +19,7 @@
     iNaT,
     to_offset,
 )
+from pandas._libs.tslibs.dtypes import NpyDatetimeUnit
 from pandas._libs.tslibs.period import INVALID_FREQ_ERR_MSG
 import pandas.util._test_decorators as td
 
@@ -338,6 +339,16 @@ def checker(res, ts, nanos):
     # --------------------------------------------------------------
     # Timestamp.replace
 
+    def test_replace_non_nano(self):
+        ts = Timestamp._from_value_and_reso(
+            91514880000000000, NpyDatetimeUnit.NPY_FR_us.value, None
+        )
+        assert ts.to_pydatetime() == datetime(4869, 12, 28)
+
+        result = ts.replace(year=4900)
+        assert result._reso == ts._reso
+        assert result.to_pydatetime() == datetime(4900, 12, 28)
+
     def test_replace_naive(self):
         # GH#14621, GH#7825
         ts = Timestamp("2016-01-01 09:00:00")