PERF: delta_to_nanoseconds (pandas-dev#47254)

Author: jbrockmendel

pandas/_libs/tslibs/dtypes.pxd

@@ -7,7 +7,7 @@ cdef str npy_unit_to_abbrev(NPY_DATETIMEUNIT unit)
 cdef NPY_DATETIMEUNIT freq_group_code_to_npy_unit(int freq) nogil
 cpdef int64_t periods_per_day(NPY_DATETIMEUNIT reso=*) except? -1
 cdef int64_t periods_per_second(NPY_DATETIMEUNIT reso) except? -1
-cdef int64_t get_conversion_factor(NPY_DATETIMEUNIT from_unit, NPY_DATETIMEUNIT to_unit)
+cdef int64_t get_conversion_factor(NPY_DATETIMEUNIT from_unit, NPY_DATETIMEUNIT to_unit) except? -1
 
 cdef dict attrname_to_abbrevs
 

pandas/_libs/tslibs/dtypes.pyx

@@ -384,7 +384,7 @@ cdef int64_t periods_per_second(NPY_DATETIMEUNIT reso) except? -1:
 
 
 @cython.overflowcheck(True)
-cdef int64_t get_conversion_factor(NPY_DATETIMEUNIT from_unit, NPY_DATETIMEUNIT to_unit):
+cdef int64_t get_conversion_factor(NPY_DATETIMEUNIT from_unit, NPY_DATETIMEUNIT to_unit) except? -1:
     """
     Find the factor by which we need to multiply to convert from from_unit to to_unit.
     """

pandas/_libs/tslibs/timedeltas.pxd

@@ -6,10 +6,11 @@ from .np_datetime cimport NPY_DATETIMEUNIT
 
 # Exposed for tslib, not intended for outside use.
 cpdef int64_t delta_to_nanoseconds(
-    delta, NPY_DATETIMEUNIT reso=*, bint round_ok=*, bint allow_year_month=*
+    delta, NPY_DATETIMEUNIT reso=*, bint round_ok=*
 ) except? -1
 cdef convert_to_timedelta64(object ts, str unit)
 cdef bint is_any_td_scalar(object obj)
+cdef object ensure_td64ns(object ts)
 
 
 cdef class _Timedelta(timedelta):

pandas/_libs/tslibs/timedeltas.pyi

@@ -76,7 +76,6 @@ def delta_to_nanoseconds(
     delta: np.timedelta64 | timedelta | Tick,
     reso: int = ...,  # NPY_DATETIMEUNIT
     round_ok: bool = ...,
-    allow_year_month: bool = ...,
 ) -> int: ...
 
 class Timedelta(timedelta):

pandas/_libs/tslibs/timedeltas.pyx

@@ -206,44 +206,24 @@ cpdef int64_t delta_to_nanoseconds(
     delta,
     NPY_DATETIMEUNIT reso=NPY_FR_ns,
     bint round_ok=True,
-    bint allow_year_month=False,
 ) except? -1:
+    # Note: this will raise on timedelta64 with Y or M unit
+
     cdef:
-        _Timedelta td
         NPY_DATETIMEUNIT in_reso
-        int64_t n
+        int64_t n, value, factor
 
     if is_tick_object(delta):
         n = delta.n
         in_reso = delta._reso
-        if in_reso == reso:
-            return n
-        else:
-            td = Timedelta._from_value_and_reso(delta.n, reso=in_reso)
 
     elif isinstance(delta, _Timedelta):
-        td = delta
         n = delta.value
         in_reso = delta._reso
-        if in_reso == reso:
-            return n
 
     elif is_timedelta64_object(delta):
         in_reso = get_datetime64_unit(delta)
         n = get_timedelta64_value(delta)
-        if in_reso == reso:
-            return n
-        else:
-            # _from_value_and_reso does not support Year, Month, or unit-less,
-            #  so we have special handling if speciifed
-            try:
-                td = Timedelta._from_value_and_reso(n, reso=in_reso)
-            except NotImplementedError:
-                if allow_year_month:
-                    td64 = ensure_td64ns(delta)
-                    return delta_to_nanoseconds(td64, reso=reso)
-                else:
-                    raise
 
     elif PyDelta_Check(delta):
         in_reso = NPY_DATETIMEUNIT.NPY_FR_us
@@ -256,21 +236,31 @@ cpdef int64_t delta_to_nanoseconds(
         except OverflowError as err:
             raise OutOfBoundsTimedelta(*err.args) from err
 
-        if in_reso == reso:
-            return n
-        else:
-            td = Timedelta._from_value_and_reso(n, reso=in_reso)
-
     else:
         raise TypeError(type(delta))
 
-    try:
-        return td._as_reso(reso, round_ok=round_ok).value
-    except OverflowError as err:
-        unit_str = npy_unit_to_abbrev(reso)
-        raise OutOfBoundsTimedelta(
-            f"Cannot cast {str(delta)} to unit={unit_str} without overflow."
-        ) from err
+    if reso < in_reso:
+        # e.g. ns -> us
+        factor = get_conversion_factor(reso, in_reso)
+        div, mod = divmod(n, factor)
+        if mod > 0 and not round_ok:
+            raise ValueError("Cannot losslessly convert units")
+
+        # Note that when mod > 0, we follow np.timedelta64 in always
+        #  rounding down.
+        value = div
+    else:
+        factor = get_conversion_factor(in_reso, reso)
+        try:
+            with cython.overflowcheck(True):
+                value = n * factor
+        except OverflowError as err:
+            unit_str = npy_unit_to_abbrev(reso)
+            raise OutOfBoundsTimedelta(
+                f"Cannot cast {str(delta)} to unit={unit_str} without overflow."
+            ) from err
+
+    return value
 
 
 @cython.overflowcheck(True)

pandas/_libs/tslibs/timestamps.pyx

@@ -99,6 +99,7 @@ from pandas._libs.tslibs.offsets cimport (
 )
 from pandas._libs.tslibs.timedeltas cimport (
     delta_to_nanoseconds,
+    ensure_td64ns,
     is_any_td_scalar,
 )
 
@@ -353,16 +354,25 @@ cdef class _Timestamp(ABCTimestamp):
             raise NotImplementedError(self._reso)
 
         if is_any_td_scalar(other):
-            if (
-                is_timedelta64_object(other)
-                and get_datetime64_unit(other) == NPY_DATETIMEUNIT.NPY_FR_GENERIC
-            ):
-                # TODO: deprecate allowing this?  We only get here
-                #  with test_timedelta_add_timestamp_interval
-                other = np.timedelta64(other.view("i8"), "ns")
-            # TODO: disallow round_ok, allow_year_month?
+            if is_timedelta64_object(other):
+                other_reso = get_datetime64_unit(other)
+                if (
+                    other_reso == NPY_DATETIMEUNIT.NPY_FR_GENERIC
+                ):
+                    # TODO: deprecate allowing this?  We only get here
+                    #  with test_timedelta_add_timestamp_interval
+                    other = np.timedelta64(other.view("i8"), "ns")
+                elif (
+                    other_reso == NPY_DATETIMEUNIT.NPY_FR_Y or other_reso == NPY_DATETIMEUNIT.NPY_FR_M
+                ):
+                    # TODO: deprecate allowing these?  or handle more like the
+                    #  corresponding DateOffsets?
+                    # TODO: no tests get here
+                    other = ensure_td64ns(other)
+
+            # TODO: disallow round_ok
             nanos = delta_to_nanoseconds(
-                other, reso=self._reso, round_ok=True, allow_year_month=True
+                other, reso=self._reso, round_ok=True
             )
             try:
                 result = type(self)(self.value + nanos, tz=self.tzinfo)

pandas/tests/scalar/timestamp/test_arithmetic.py

@@ -45,12 +45,6 @@ def test_overflow_offset_raises(self):
             r"\<-?\d+ \* Days\> and \d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2} "
             "will overflow"
         )
-        lmsg = "|".join(
-            [
-                "Python int too large to convert to C (long|int)",
-                "int too big to convert",
-            ]
-        )
         lmsg2 = r"Cannot cast <-?20169940 \* Days> to unit=ns without overflow"
 
         with pytest.raises(OutOfBoundsTimedelta, match=lmsg2):
@@ -68,13 +62,14 @@ def test_overflow_offset_raises(self):
         stamp = Timestamp("2000/1/1")
         offset_overflow = to_offset("D") * 100**5
 
-        with pytest.raises(OverflowError, match=lmsg):
+        lmsg3 = r"Cannot cast <-?10000000000 \* Days> to unit=ns without overflow"
+        with pytest.raises(OutOfBoundsTimedelta, match=lmsg3):
             stamp + offset_overflow
 
         with pytest.raises(OverflowError, match=msg):
             offset_overflow + stamp
 
-        with pytest.raises(OverflowError, match=lmsg):
+        with pytest.raises(OutOfBoundsTimedelta, match=lmsg3):
             stamp - offset_overflow
 
     def test_overflow_timestamp_raises(self):

pandas/tests/tslibs/test_timedeltas.py

@@ -55,6 +55,18 @@ def test_delta_to_nanoseconds_error():
         delta_to_nanoseconds(np.int32(3))
 
 
+def test_delta_to_nanoseconds_td64_MY_raises():
+    td = np.timedelta64(1234, "Y")
+
+    with pytest.raises(ValueError, match="0, 10"):
+        delta_to_nanoseconds(td)
+
+    td = np.timedelta64(1234, "M")
+
+    with pytest.raises(ValueError, match="1, 10"):
+        delta_to_nanoseconds(td)
+
+
 def test_huge_nanoseconds_overflow():
     # GH 32402
     assert delta_to_nanoseconds(Timedelta(1e10)) == 1e10