Merge branch 'master' of https://github.com/pandas-dev/pandas into ref-liboffsets-cdefit

Author: jbrockmendel

doc/source/user_guide/computation.rst

@@ -648,6 +648,24 @@ from present information back to past information. This allows the rolling windo
 Currently, this feature is only implemented for time-based windows.
 For fixed windows, the closed parameter cannot be set and the rolling window will always have both endpoints closed.
 
+.. _stats.iter_rolling_window:
+
+Iteration over window:
+~~~~~~~~~~~~~~~~~~~~~~
+
+.. versionadded:: 1.1.0
+
+``Rolling`` and ``Expanding`` objects now support iteration. Be noted that ``min_periods`` is ignored in iteration.
+
+.. ipython::
+
+   In [1]: df = pd.DataFrame({"A": [1, 2, 3], "B": [4, 5, 6]})
+
+   In [2]: for i in df.rolling(2):
+      ...:     print(i)
+      ...:
+
+
 .. _stats.moments.ts-versus-resampling:
 
 Time-aware rolling vs. resampling

doc/source/whatsnew/v1.1.0.rst

@@ -235,6 +235,7 @@ Other enhancements
   :class:`~pandas.io.stata.StataWriter`, :class:`~pandas.io.stata.StataWriter117`,
   and :class:`~pandas.io.stata.StataWriterUTF8` (:issue:`26599`).
 - :meth:`HDFStore.put` now accepts `track_times` parameter. Parameter is passed to ``create_table`` method of ``PyTables`` (:issue:`32682`).
+- Make :class:`pandas.core.window.Rolling` and :class:`pandas.core.window.Expanding` iterable（:issue:`11704`)
 
 .. ---------------------------------------------------------------------------
 

pandas/_libs/tslibs/frequencies.pxd

@@ -1,7 +1,6 @@
 cpdef str get_rule_month(object source, str default=*)
 
 cpdef get_freq_code(freqstr)
-cpdef object get_freq(object freq)
 cpdef str get_base_alias(freqstr)
 cpdef int get_to_timestamp_base(int base)
 cpdef str get_freq_str(base, mult=*)

pandas/_libs/tslibs/frequencies.pyx

@@ -306,25 +306,6 @@ cpdef int get_to_timestamp_base(int base):
     return base
 
 
-cpdef object get_freq(object freq):
-    """
-    Return frequency code of given frequency str.
-    If input is not string, return input as it is.
-
-    Examples
-    --------
-    >>> get_freq('A')
-    1000
-
-    >>> get_freq('3A')
-    1000
-    """
-    if isinstance(freq, str):
-        base, mult = get_freq_code(freq)
-        freq = base
-    return freq
-
-
 # ----------------------------------------------------------------------
 # Frequency comparison
 

pandas/_libs/tslibs/np_datetime.pxd

@@ -53,8 +53,6 @@ cdef extern from "src/datetime/np_datetime.h":
                                            npy_datetimestruct *result) nogil
 
 
-cdef int reverse_ops[6]
-
 cdef bint cmp_scalar(int64_t lhs, int64_t rhs, int op) except -1
 
 cdef check_dts_bounds(npy_datetimestruct *dts)

pandas/_libs/tslibs/np_datetime.pyx

@@ -68,15 +68,6 @@ cdef inline NPY_DATETIMEUNIT get_datetime64_unit(object obj) nogil:
 # ----------------------------------------------------------------------
 # Comparison
 
-cdef int reverse_ops[6]
-
-reverse_ops[Py_LT] = Py_GT
-reverse_ops[Py_LE] = Py_GE
-reverse_ops[Py_EQ] = Py_EQ
-reverse_ops[Py_NE] = Py_NE
-reverse_ops[Py_GT] = Py_LT
-reverse_ops[Py_GE] = Py_LE
-
 
 cdef inline bint cmp_scalar(int64_t lhs, int64_t rhs, int op) except -1:
     """

pandas/_libs/tslibs/offsets.pyx

@@ -23,7 +23,7 @@ from pandas._libs.tslibs.util cimport is_integer_object, is_datetime64_object
 
 from pandas._libs.tslibs.base cimport ABCTimestamp
 
-from pandas._libs.tslibs.ccalendar import MONTHS, DAYS
+from pandas._libs.tslibs.ccalendar import MONTHS, DAYS, weekday_to_int, int_to_weekday
 from pandas._libs.tslibs.ccalendar cimport get_days_in_month, dayofweek
 from pandas._libs.tslibs.conversion cimport (
     convert_datetime_to_tsobject,
@@ -993,10 +993,16 @@ class CustomMixin:
         object.__setattr__(self, "calendar", calendar)
 
 
-class WeekOfMonthMixin:
+class WeekOfMonthMixin(BaseOffset):
     """
     Mixin for methods common to WeekOfMonth and LastWeekOfMonth.
     """
+    def __init__(self, n=1, normalize=False, weekday=0):
+        BaseOffset.__init__(self, n, normalize)
+        object.__setattr__(self, "weekday", weekday)
+
+        if weekday < 0 or weekday > 6:
+            raise ValueError(f"Day must be 0<=day<=6, got {weekday}")
 
     @apply_wraps
     def apply(self, other):
@@ -1017,6 +1023,14 @@ class WeekOfMonthMixin:
             return False
         return dt.day == self._get_offset_day(dt)
 
+    @property
+    def rule_code(self) -> str:
+        weekday = int_to_weekday.get(self.weekday, "")
+        if self.week == -1:
+            # LastWeekOfMonth
+            return f"{self._prefix}-{weekday}"
+        return f"{self._prefix}-{self.week + 1}{weekday}"
+
 
 # ----------------------------------------------------------------------
 # RelativeDelta Arithmetic

pandas/_libs/tslibs/timedeltas.pyx

@@ -26,7 +26,7 @@ from pandas._libs.tslibs.base cimport ABCTimedelta, ABCTimestamp
 from pandas._libs.tslibs.ccalendar cimport DAY_NANOS
 
 from pandas._libs.tslibs.np_datetime cimport (
-    cmp_scalar, reverse_ops, td64_to_tdstruct, pandas_timedeltastruct)
+    cmp_scalar, td64_to_tdstruct, pandas_timedeltastruct)
 
 from pandas._libs.tslibs.nattype cimport (
     checknull_with_nat,

pandas/_libs/tslibs/timestamps.pxd

@@ -15,7 +15,7 @@ cdef class _Timestamp(ABCTimestamp):
     cdef readonly:
         int64_t value, nanosecond
         object freq
-        list _date_attributes
+
     cpdef bint _get_start_end_field(self, str field)
     cpdef _get_date_name_field(self, object field, object locale)
     cdef int64_t _maybe_convert_value_to_local(self)

pandas/_libs/tslibs/timestamps.pyx

@@ -387,6 +387,10 @@ cdef class _Timestamp(ABCTimestamp):
                     dtype=object,
                 )
 
+        elif not isinstance(self, _Timestamp):
+            # cython semantics, args have been switched and this is __radd__
+            return other.__add__(self)
+
         return NotImplemented
 
     def __sub__(self, other):
@@ -1051,7 +1055,7 @@ timedelta}, default 'raise'
         return Period(self, freq=freq)
 
     @property
-    def dayofweek(self):
+    def dayofweek(self) -> int:
         """
         Return day of the week.
         """
@@ -1092,7 +1096,7 @@ timedelta}, default 'raise'
         return self._get_date_name_field('month_name', locale)
 
     @property
-    def dayofyear(self):
+    def dayofyear(self) -> int:
         """
         Return the day of the year.
         """
@@ -1115,7 +1119,7 @@ timedelta}, default 'raise'
         return ((self.month - 1) // 3) + 1
 
     @property
-    def days_in_month(self):
+    def days_in_month(self) -> int:
         """
         Return the number of days in the month.
         """
@@ -1428,16 +1432,7 @@ default 'raise'
 
         return base1 + base2
 
-    def _has_time_component(self) -> bool:
-        """
-        Returns if the Timestamp has a time component
-        in addition to the date part
-        """
-        return (self.time() != _zero_time
-                or self.tzinfo is not None
-                or self.nanosecond != 0)
-
-    def to_julian_date(self):
+    def to_julian_date(self) -> np.float64:
         """
         Convert TimeStamp to a Julian Date.
         0 Julian date is noon January 1, 4713 BC.
@@ -1474,11 +1469,6 @@ default 'raise'
             np.array([self.value], dtype='i8'), tz=self.tz)[0]
         return Timestamp(normalized_value).tz_localize(self.tz)
 
-    def __radd__(self, other):
-        # __radd__ on cython extension types like _Timestamp is not used, so
-        # define it here instead
-        return self + other
-
 
 # Add the min and max fields at the class level
 cdef int64_t _NS_UPPER_BOUND = np.iinfo(np.int64).max

pandas/core/arrays/categorical.py

@@ -64,6 +64,10 @@ def _cat_compare_op(op):
 
     @unpack_zerodim_and_defer(opname)
     def func(self, other):
+        if is_list_like(other) and len(other) != len(self):
+            # TODO: Could this fail if the categories are listlike objects?
+            raise ValueError("Lengths must match.")
+
         if not self.ordered:
             if opname in ["__lt__", "__gt__", "__le__", "__ge__"]:
                 raise TypeError(

pandas/core/arrays/datetimelike.py

@@ -84,6 +84,9 @@ def _validate_comparison_value(self, other):
         elif not is_list_like(other):
             raise InvalidComparison(other)
 
+        elif len(other) != len(self):
+            raise ValueError("Lengths must match")
+
         else:
             try:
                 other = self._validate_listlike(other, opname, allow_object=True)
@@ -1234,6 +1237,9 @@ def _add_timedelta_arraylike(self, other):
         """
         # overridden by PeriodArray
 
+        if len(self) != len(other):
+            raise ValueError("cannot add indices of unequal length")
+
         if isinstance(other, np.ndarray):
             # ndarray[timedelta64]; wrap in TimedeltaIndex for op
             from pandas.core.arrays import TimedeltaArray

pandas/core/arrays/datetimes.py

@@ -654,6 +654,9 @@ def _assert_tzawareness_compat(self, other):
 
     def _sub_datetime_arraylike(self, other):
         """subtract DatetimeArray/Index or ndarray[datetime64]"""
+        if len(self) != len(other):
+            raise ValueError("cannot add indices of unequal length")
+
         if isinstance(other, np.ndarray):
             assert is_datetime64_dtype(other)
             other = type(self)(other)

pandas/core/arrays/integer.py

@@ -517,6 +517,8 @@ def cmp_method(self, other):
                     raise NotImplementedError(
                         "can only perform ops with 1-d structures"
                     )
+                if len(self) != len(other):
+                    raise ValueError("Lengths must match to compare")
 
             if other is libmissing.NA:
                 # numpy does not handle pd.NA well as "other" scalar (it returns
@@ -620,6 +622,8 @@ def integer_arithmetic_method(self, other):
                     raise NotImplementedError(
                         "can only perform ops with 1-d structures"
                     )
+                if len(self) != len(other):
+                    raise ValueError("Lengths must match")
                 if not (is_float_dtype(other) or is_integer_dtype(other)):
                     raise TypeError("can only perform ops with numeric values")
 

pandas/core/arrays/period.py

@@ -746,8 +746,7 @@ def _check_timedeltalike_freq_compat(self, other):
         IncompatibleFrequency
         """
         assert isinstance(self.freq, Tick)  # checked by calling function
-        own_offset = frequencies.to_offset(self.freq.rule_code)
-        base_nanos = delta_to_nanoseconds(own_offset)
+        base_nanos = self.freq.base.nanos
 
         if isinstance(other, (timedelta, np.timedelta64, Tick)):
             nanos = delta_to_nanoseconds(other)

pandas/core/arrays/timedeltas.py

@@ -471,6 +471,10 @@ def __mul__(self, other):
         if not hasattr(other, "dtype"):
             # list, tuple
             other = np.array(other)
+        if len(other) != len(self) and not is_timedelta64_dtype(other.dtype):
+            # Exclude timedelta64 here so we correctly raise TypeError
+            #  for that instead of ValueError
+            raise ValueError("Cannot multiply with unequal lengths")
 
         if is_object_dtype(other.dtype):
             # this multiplication will succeed only if all elements of other
@@ -514,7 +518,10 @@ def __truediv__(self, other):
             # e.g. list, tuple
             other = np.array(other)
 
-        if is_timedelta64_dtype(other.dtype):
+        if len(other) != len(self):
+            raise ValueError("Cannot divide vectors with unequal lengths")
+
+        elif is_timedelta64_dtype(other.dtype):
             # let numpy handle it
             return self._data / other
 
@@ -564,7 +571,10 @@ def __rtruediv__(self, other):
             # e.g. list, tuple
             other = np.array(other)
 
-        if is_timedelta64_dtype(other.dtype):
+        if len(other) != len(self):
+            raise ValueError("Cannot divide vectors with unequal lengths")
+
+        elif is_timedelta64_dtype(other.dtype):
             # let numpy handle it
             return other / self._data
 
@@ -613,8 +623,10 @@ def __floordiv__(self, other):
         if not hasattr(other, "dtype"):
             # list, tuple
             other = np.array(other)
+        if len(other) != len(self):
+            raise ValueError("Cannot divide with unequal lengths")
 
-        if is_timedelta64_dtype(other.dtype):
+        elif is_timedelta64_dtype(other.dtype):
             other = type(self)(other)
 
             # numpy timedelta64 does not natively support floordiv, so operate
@@ -666,7 +678,10 @@ def __rfloordiv__(self, other):
             # list, tuple
             other = np.array(other)
 
-        if is_timedelta64_dtype(other.dtype):
+        if len(other) != len(self):
+            raise ValueError("Cannot divide with unequal lengths")
+
+        elif is_timedelta64_dtype(other.dtype):
             other = type(self)(other)
 
             # numpy timedelta64 does not natively support floordiv, so operate