manyi/venv/Lib/site-packages/pandas/tests/arithmetic/test_timedelta64.py

# -*- coding: utf-8 -*-
# Arithmetc tests for DataFrame/Series/Index/Array classes that should
# behave identically.
from datetime import datetime, timedelta

import numpy as np
import pytest

from pandas.errors import NullFrequencyError, PerformanceWarning

import pandas as pd
from pandas import (
    DataFrame, DatetimeIndex, NaT, Series, Timedelta, TimedeltaIndex,
    Timestamp, timedelta_range)
import pandas.util.testing as tm


def get_upcast_box(box, vector):
    """
    Given two box-types, find the one that takes priority
    """
    if box is DataFrame or isinstance(vector, DataFrame):
        return DataFrame
    if box is Series or isinstance(vector, Series):
        return Series
    if box is pd.Index or isinstance(vector, pd.Index):
        return pd.Index
    return box


# ------------------------------------------------------------------
# Timedelta64[ns] dtype Comparisons

class TestTimedelta64ArrayComparisons(object):
    # TODO: All of these need to be parametrized over box

    def test_compare_timedelta_series(self):
        # regresssion test for GH#5963
        s = pd.Series([timedelta(days=1), timedelta(days=2)])
        actual = s > timedelta(days=1)
        expected = pd.Series([False, True])
        tm.assert_series_equal(actual, expected)

    def test_tdi_cmp_str_invalid(self, box_with_array):
        # GH#13624
        xbox = box_with_array if box_with_array is not pd.Index else np.ndarray
        tdi = TimedeltaIndex(['1 day', '2 days'])
        tdarr = tm.box_expected(tdi, box_with_array)

        for left, right in [(tdarr, 'a'), ('a', tdarr)]:
            with pytest.raises(TypeError):
                left > right
            with pytest.raises(TypeError):
                left >= right
            with pytest.raises(TypeError):
                left < right
            with pytest.raises(TypeError):
                left <= right

            result = left == right
            expected = np.array([False, False], dtype=bool)
            expected = tm.box_expected(expected, xbox)
            tm.assert_equal(result, expected)

            result = left != right
            expected = np.array([True, True], dtype=bool)
            expected = tm.box_expected(expected, xbox)
            tm.assert_equal(result, expected)

    @pytest.mark.parametrize('dtype', [None, object])
    def test_comp_nat(self, dtype):
        left = pd.TimedeltaIndex([pd.Timedelta('1 days'), pd.NaT,
                                  pd.Timedelta('3 days')])
        right = pd.TimedeltaIndex([pd.NaT, pd.NaT, pd.Timedelta('3 days')])

        lhs, rhs = left, right
        if dtype is object:
            lhs, rhs = left.astype(object), right.astype(object)

        result = rhs == lhs
        expected = np.array([False, False, True])
        tm.assert_numpy_array_equal(result, expected)

        result = rhs != lhs
        expected = np.array([True, True, False])
        tm.assert_numpy_array_equal(result, expected)

        expected = np.array([False, False, False])
        tm.assert_numpy_array_equal(lhs == pd.NaT, expected)
        tm.assert_numpy_array_equal(pd.NaT == rhs, expected)

        expected = np.array([True, True, True])
        tm.assert_numpy_array_equal(lhs != pd.NaT, expected)
        tm.assert_numpy_array_equal(pd.NaT != lhs, expected)

        expected = np.array([False, False, False])
        tm.assert_numpy_array_equal(lhs < pd.NaT, expected)
        tm.assert_numpy_array_equal(pd.NaT > lhs, expected)

    def test_comparisons_nat(self):
        tdidx1 = pd.TimedeltaIndex(['1 day', pd.NaT, '1 day 00:00:01', pd.NaT,
                                    '1 day 00:00:01', '5 day 00:00:03'])
        tdidx2 = pd.TimedeltaIndex(['2 day', '2 day', pd.NaT, pd.NaT,
                                    '1 day 00:00:02', '5 days 00:00:03'])
        tdarr = np.array([np.timedelta64(2, 'D'),
                          np.timedelta64(2, 'D'), np.timedelta64('nat'),
                          np.timedelta64('nat'),
                          np.timedelta64(1, 'D') + np.timedelta64(2, 's'),
                          np.timedelta64(5, 'D') + np.timedelta64(3, 's')])

        cases = [(tdidx1, tdidx2), (tdidx1, tdarr)]

        # Check pd.NaT is handles as the same as np.nan
        for idx1, idx2 in cases:

            result = idx1 < idx2
            expected = np.array([True, False, False, False, True, False])
            tm.assert_numpy_array_equal(result, expected)

            result = idx2 > idx1
            expected = np.array([True, False, False, False, True, False])
            tm.assert_numpy_array_equal(result, expected)

            result = idx1 <= idx2
            expected = np.array([True, False, False, False, True, True])
            tm.assert_numpy_array_equal(result, expected)

            result = idx2 >= idx1
            expected = np.array([True, False, False, False, True, True])
            tm.assert_numpy_array_equal(result, expected)

            result = idx1 == idx2
            expected = np.array([False, False, False, False, False, True])
            tm.assert_numpy_array_equal(result, expected)

            result = idx1 != idx2
            expected = np.array([True, True, True, True, True, False])
            tm.assert_numpy_array_equal(result, expected)

    # TODO: better name
    def test_comparisons_coverage(self):
        rng = timedelta_range('1 days', periods=10)

        result = rng < rng[3]
        expected = np.array([True, True, True] + [False] * 7)
        tm.assert_numpy_array_equal(result, expected)

        # raise TypeError for now
        with pytest.raises(TypeError):
            rng < rng[3].value

        result = rng == list(rng)
        exp = rng == rng
        tm.assert_numpy_array_equal(result, exp)


# ------------------------------------------------------------------
# Timedelta64[ns] dtype Arithmetic Operations

class TestTimedelta64ArithmeticUnsorted(object):
    # Tests moved from type-specific test files but not
    #  yet sorted/parametrized/de-duplicated

    def test_ufunc_coercions(self):
        # normal ops are also tested in tseries/test_timedeltas.py
        idx = TimedeltaIndex(['2H', '4H', '6H', '8H', '10H'],
                             freq='2H', name='x')

        for result in [idx * 2, np.multiply(idx, 2)]:
            assert isinstance(result, TimedeltaIndex)
            exp = TimedeltaIndex(['4H', '8H', '12H', '16H', '20H'],
                                 freq='4H', name='x')
            tm.assert_index_equal(result, exp)
            assert result.freq == '4H'

        for result in [idx / 2, np.divide(idx, 2)]:
            assert isinstance(result, TimedeltaIndex)
            exp = TimedeltaIndex(['1H', '2H', '3H', '4H', '5H'],
                                 freq='H', name='x')
            tm.assert_index_equal(result, exp)
            assert result.freq == 'H'

        idx = TimedeltaIndex(['2H', '4H', '6H', '8H', '10H'],
                             freq='2H', name='x')
        for result in [-idx, np.negative(idx)]:
            assert isinstance(result, TimedeltaIndex)
            exp = TimedeltaIndex(['-2H', '-4H', '-6H', '-8H', '-10H'],
                                 freq='-2H', name='x')
            tm.assert_index_equal(result, exp)
            assert result.freq == '-2H'

        idx = TimedeltaIndex(['-2H', '-1H', '0H', '1H', '2H'],
                             freq='H', name='x')
        for result in [abs(idx), np.absolute(idx)]:
            assert isinstance(result, TimedeltaIndex)
            exp = TimedeltaIndex(['2H', '1H', '0H', '1H', '2H'],
                                 freq=None, name='x')
            tm.assert_index_equal(result, exp)
            assert result.freq is None

    def test_subtraction_ops(self):
        # with datetimes/timedelta and tdi/dti
        tdi = TimedeltaIndex(['1 days', pd.NaT, '2 days'], name='foo')
        dti = pd.date_range('20130101', periods=3, name='bar')
        td = Timedelta('1 days')
        dt = Timestamp('20130101')

        pytest.raises(TypeError, lambda: tdi - dt)
        pytest.raises(TypeError, lambda: tdi - dti)
        pytest.raises(TypeError, lambda: td - dt)
        pytest.raises(TypeError, lambda: td - dti)

        result = dt - dti
        expected = TimedeltaIndex(['0 days', '-1 days', '-2 days'], name='bar')
        tm.assert_index_equal(result, expected)

        result = dti - dt
        expected = TimedeltaIndex(['0 days', '1 days', '2 days'], name='bar')
        tm.assert_index_equal(result, expected)

        result = tdi - td
        expected = TimedeltaIndex(['0 days', pd.NaT, '1 days'], name='foo')
        tm.assert_index_equal(result, expected, check_names=False)

        result = td - tdi
        expected = TimedeltaIndex(['0 days', pd.NaT, '-1 days'], name='foo')
        tm.assert_index_equal(result, expected, check_names=False)

        result = dti - td
        expected = DatetimeIndex(
            ['20121231', '20130101', '20130102'], name='bar')
        tm.assert_index_equal(result, expected, check_names=False)

        result = dt - tdi
        expected = DatetimeIndex(['20121231', pd.NaT, '20121230'], name='foo')
        tm.assert_index_equal(result, expected)

    def test_subtraction_ops_with_tz(self):

        # check that dt/dti subtraction ops with tz are validated
        dti = pd.date_range('20130101', periods=3)
        ts = Timestamp('20130101')
        dt = ts.to_pydatetime()
        dti_tz = pd.date_range('20130101', periods=3).tz_localize('US/Eastern')
        ts_tz = Timestamp('20130101').tz_localize('US/Eastern')
        ts_tz2 = Timestamp('20130101').tz_localize('CET')
        dt_tz = ts_tz.to_pydatetime()
        td = Timedelta('1 days')

        def _check(result, expected):
            assert result == expected
            assert isinstance(result, Timedelta)

        # scalars
        result = ts - ts
        expected = Timedelta('0 days')
        _check(result, expected)

        result = dt_tz - ts_tz
        expected = Timedelta('0 days')
        _check(result, expected)

        result = ts_tz - dt_tz
        expected = Timedelta('0 days')
        _check(result, expected)

        # tz mismatches
        pytest.raises(TypeError, lambda: dt_tz - ts)
        pytest.raises(TypeError, lambda: dt_tz - dt)
        pytest.raises(TypeError, lambda: dt_tz - ts_tz2)
        pytest.raises(TypeError, lambda: dt - dt_tz)
        pytest.raises(TypeError, lambda: ts - dt_tz)
        pytest.raises(TypeError, lambda: ts_tz2 - ts)
        pytest.raises(TypeError, lambda: ts_tz2 - dt)
        pytest.raises(TypeError, lambda: ts_tz - ts_tz2)

        # with dti
        pytest.raises(TypeError, lambda: dti - ts_tz)
        pytest.raises(TypeError, lambda: dti_tz - ts)
        pytest.raises(TypeError, lambda: dti_tz - ts_tz2)

        result = dti_tz - dt_tz
        expected = TimedeltaIndex(['0 days', '1 days', '2 days'])
        tm.assert_index_equal(result, expected)

        result = dt_tz - dti_tz
        expected = TimedeltaIndex(['0 days', '-1 days', '-2 days'])
        tm.assert_index_equal(result, expected)

        result = dti_tz - ts_tz
        expected = TimedeltaIndex(['0 days', '1 days', '2 days'])
        tm.assert_index_equal(result, expected)

        result = ts_tz - dti_tz
        expected = TimedeltaIndex(['0 days', '-1 days', '-2 days'])
        tm.assert_index_equal(result, expected)

        result = td - td
        expected = Timedelta('0 days')
        _check(result, expected)

        result = dti_tz - td
        expected = DatetimeIndex(
            ['20121231', '20130101', '20130102'], tz='US/Eastern')
        tm.assert_index_equal(result, expected)

    def test_dti_tdi_numeric_ops(self):
        # These are normally union/diff set-like ops
        tdi = TimedeltaIndex(['1 days', pd.NaT, '2 days'], name='foo')
        dti = pd.date_range('20130101', periods=3, name='bar')

        # TODO(wesm): unused?
        # td = Timedelta('1 days')
        # dt = Timestamp('20130101')

        result = tdi - tdi
        expected = TimedeltaIndex(['0 days', pd.NaT, '0 days'], name='foo')
        tm.assert_index_equal(result, expected)

        result = tdi + tdi
        expected = TimedeltaIndex(['2 days', pd.NaT, '4 days'], name='foo')
        tm.assert_index_equal(result, expected)

        result = dti - tdi  # name will be reset
        expected = DatetimeIndex(['20121231', pd.NaT, '20130101'])
        tm.assert_index_equal(result, expected)

    def test_addition_ops(self):
        # with datetimes/timedelta and tdi/dti
        tdi = TimedeltaIndex(['1 days', pd.NaT, '2 days'], name='foo')
        dti = pd.date_range('20130101', periods=3, name='bar')
        td = Timedelta('1 days')
        dt = Timestamp('20130101')

        result = tdi + dt
        expected = DatetimeIndex(['20130102', pd.NaT, '20130103'], name='foo')
        tm.assert_index_equal(result, expected)

        result = dt + tdi
        expected = DatetimeIndex(['20130102', pd.NaT, '20130103'], name='foo')
        tm.assert_index_equal(result, expected)

        result = td + tdi
        expected = TimedeltaIndex(['2 days', pd.NaT, '3 days'], name='foo')
        tm.assert_index_equal(result, expected)

        result = tdi + td
        expected = TimedeltaIndex(['2 days', pd.NaT, '3 days'], name='foo')
        tm.assert_index_equal(result, expected)

        # unequal length
        pytest.raises(ValueError, lambda: tdi + dti[0:1])
        pytest.raises(ValueError, lambda: tdi[0:1] + dti)

        # random indexes
        with pytest.raises(NullFrequencyError):
            tdi + pd.Int64Index([1, 2, 3])

        # this is a union!
        # pytest.raises(TypeError, lambda : Int64Index([1,2,3]) + tdi)

        result = tdi + dti  # name will be reset
        expected = DatetimeIndex(['20130102', pd.NaT, '20130105'])
        tm.assert_index_equal(result, expected)

        result = dti + tdi  # name will be reset
        expected = DatetimeIndex(['20130102', pd.NaT, '20130105'])
        tm.assert_index_equal(result, expected)

        result = dt + td
        expected = Timestamp('20130102')
        assert result == expected

        result = td + dt
        expected = Timestamp('20130102')
        assert result == expected

    # TODO: Needs more informative name, probably split up into
    # more targeted tests
    @pytest.mark.parametrize('freq', ['D', 'B'])
    def test_timedelta(self, freq):
        index = pd.date_range('1/1/2000', periods=50, freq=freq)

        shifted = index + timedelta(1)
        back = shifted + timedelta(-1)
        tm.assert_index_equal(index, back)

        if freq == 'D':
            expected = pd.tseries.offsets.Day(1)
            assert index.freq == expected
            assert shifted.freq == expected
            assert back.freq == expected
        else:  # freq == 'B'
            assert index.freq == pd.tseries.offsets.BusinessDay(1)
            assert shifted.freq is None
            assert back.freq == pd.tseries.offsets.BusinessDay(1)

        result = index - timedelta(1)
        expected = index + timedelta(-1)
        tm.assert_index_equal(result, expected)

        # GH#4134, buggy with timedeltas
        rng = pd.date_range('2013', '2014')
        s = Series(rng)
        result1 = rng - pd.offsets.Hour(1)
        result2 = DatetimeIndex(s - np.timedelta64(100000000))
        result3 = rng - np.timedelta64(100000000)
        result4 = DatetimeIndex(s - pd.offsets.Hour(1))
        tm.assert_index_equal(result1, result4)
        tm.assert_index_equal(result2, result3)


class TestAddSubNaTMasking(object):
    # TODO: parametrize over boxes

    def test_tdi_add_timestamp_nat_masking(self):
        # GH#17991 checking for overflow-masking with NaT
        tdinat = pd.to_timedelta(['24658 days 11:15:00', 'NaT'])

        tsneg = Timestamp('1950-01-01')
        ts_neg_variants = [tsneg,
                           tsneg.to_pydatetime(),
                           tsneg.to_datetime64().astype('datetime64[ns]'),
                           tsneg.to_datetime64().astype('datetime64[D]')]

        tspos = Timestamp('1980-01-01')
        ts_pos_variants = [tspos,
                           tspos.to_pydatetime(),
                           tspos.to_datetime64().astype('datetime64[ns]'),
                           tspos.to_datetime64().astype('datetime64[D]')]

        for variant in ts_neg_variants + ts_pos_variants:
            res = tdinat + variant
            assert res[1] is pd.NaT

    def test_tdi_add_overflow(self):
        # See GH#14068
        msg = "too (big|large) to convert"
        with pytest.raises(OverflowError, match=msg):
            pd.to_timedelta(106580, 'D') + Timestamp('2000')
        with pytest.raises(OverflowError, match=msg):
            Timestamp('2000') + pd.to_timedelta(106580, 'D')

        _NaT = int(pd.NaT) + 1
        msg = "Overflow in int64 addition"
        with pytest.raises(OverflowError, match=msg):
            pd.to_timedelta([106580], 'D') + Timestamp('2000')
        with pytest.raises(OverflowError, match=msg):
            Timestamp('2000') + pd.to_timedelta([106580], 'D')
        with pytest.raises(OverflowError, match=msg):
            pd.to_timedelta([_NaT]) - Timedelta('1 days')
        with pytest.raises(OverflowError, match=msg):
            pd.to_timedelta(['5 days', _NaT]) - Timedelta('1 days')
        with pytest.raises(OverflowError, match=msg):
            (pd.to_timedelta([_NaT, '5 days', '1 hours']) -
             pd.to_timedelta(['7 seconds', _NaT, '4 hours']))

        # These should not overflow!
        exp = TimedeltaIndex([pd.NaT])
        result = pd.to_timedelta([pd.NaT]) - Timedelta('1 days')
        tm.assert_index_equal(result, exp)

        exp = TimedeltaIndex(['4 days', pd.NaT])
        result = pd.to_timedelta(['5 days', pd.NaT]) - Timedelta('1 days')
        tm.assert_index_equal(result, exp)

        exp = TimedeltaIndex([pd.NaT, pd.NaT, '5 hours'])
        result = (pd.to_timedelta([pd.NaT, '5 days', '1 hours']) +
                  pd.to_timedelta(['7 seconds', pd.NaT, '4 hours']))
        tm.assert_index_equal(result, exp)


class TestTimedeltaArraylikeAddSubOps(object):
    # Tests for timedelta64[ns] __add__, __sub__, __radd__, __rsub__

    # TODO: moved from frame tests; needs parametrization/de-duplication
    def test_td64_df_add_int_frame(self):
        # GH#22696 Check that we don't dispatch to numpy implementation,
        #  which treats int64 as m8[ns]
        tdi = pd.timedelta_range('1', periods=3)
        df = tdi.to_frame()
        other = pd.DataFrame([1, 2, 3], index=tdi)  # indexed like `df`
        with pytest.raises(TypeError):
            df + other
        with pytest.raises(TypeError):
            other + df
        with pytest.raises(TypeError):
            df - other
        with pytest.raises(TypeError):
            other - df

    # TODO: moved from tests.indexes.timedeltas.test_arithmetic; needs
    #  parametrization+de-duplication
    def test_timedelta_ops_with_missing_values(self):
        # setup
        s1 = pd.to_timedelta(Series(['00:00:01']))
        s2 = pd.to_timedelta(Series(['00:00:02']))
        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
            # Passing datetime64-dtype data to TimedeltaIndex is deprecated
            sn = pd.to_timedelta(Series([pd.NaT]))

        df1 = pd.DataFrame(['00:00:01']).apply(pd.to_timedelta)
        df2 = pd.DataFrame(['00:00:02']).apply(pd.to_timedelta)
        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
            # Passing datetime64-dtype data to TimedeltaIndex is deprecated
            dfn = pd.DataFrame([pd.NaT]).apply(pd.to_timedelta)

        scalar1 = pd.to_timedelta('00:00:01')
        scalar2 = pd.to_timedelta('00:00:02')
        timedelta_NaT = pd.to_timedelta('NaT')

        actual = scalar1 + scalar1
        assert actual == scalar2
        actual = scalar2 - scalar1
        assert actual == scalar1

        actual = s1 + s1
        tm.assert_series_equal(actual, s2)
        actual = s2 - s1
        tm.assert_series_equal(actual, s1)

        actual = s1 + scalar1
        tm.assert_series_equal(actual, s2)
        actual = scalar1 + s1
        tm.assert_series_equal(actual, s2)
        actual = s2 - scalar1
        tm.assert_series_equal(actual, s1)
        actual = -scalar1 + s2
        tm.assert_series_equal(actual, s1)

        actual = s1 + timedelta_NaT
        tm.assert_series_equal(actual, sn)
        actual = timedelta_NaT + s1
        tm.assert_series_equal(actual, sn)
        actual = s1 - timedelta_NaT
        tm.assert_series_equal(actual, sn)
        actual = -timedelta_NaT + s1
        tm.assert_series_equal(actual, sn)

        with pytest.raises(TypeError):
            s1 + np.nan
        with pytest.raises(TypeError):
            np.nan + s1
        with pytest.raises(TypeError):
            s1 - np.nan
        with pytest.raises(TypeError):
            -np.nan + s1

        actual = s1 + pd.NaT
        tm.assert_series_equal(actual, sn)
        actual = s2 - pd.NaT
        tm.assert_series_equal(actual, sn)

        actual = s1 + df1
        tm.assert_frame_equal(actual, df2)
        actual = s2 - df1
        tm.assert_frame_equal(actual, df1)
        actual = df1 + s1
        tm.assert_frame_equal(actual, df2)
        actual = df2 - s1
        tm.assert_frame_equal(actual, df1)

        actual = df1 + df1
        tm.assert_frame_equal(actual, df2)
        actual = df2 - df1
        tm.assert_frame_equal(actual, df1)

        actual = df1 + scalar1
        tm.assert_frame_equal(actual, df2)
        actual = df2 - scalar1
        tm.assert_frame_equal(actual, df1)

        actual = df1 + timedelta_NaT
        tm.assert_frame_equal(actual, dfn)
        actual = df1 - timedelta_NaT
        tm.assert_frame_equal(actual, dfn)

        with pytest.raises(TypeError):
            df1 + np.nan
        with pytest.raises(TypeError):
            df1 - np.nan

        actual = df1 + pd.NaT  # NaT is datetime, not timedelta
        tm.assert_frame_equal(actual, dfn)
        actual = df1 - pd.NaT
        tm.assert_frame_equal(actual, dfn)

    # TODO: moved from tests.series.test_operators, needs splitting, cleanup,
    # de-duplication, box-parametrization...
    def test_operators_timedelta64(self):
        # series ops
        v1 = pd.date_range('2012-1-1', periods=3, freq='D')
        v2 = pd.date_range('2012-1-2', periods=3, freq='D')
        rs = Series(v2) - Series(v1)
        xp = Series(1e9 * 3600 * 24,
                    rs.index).astype('int64').astype('timedelta64[ns]')
        tm.assert_series_equal(rs, xp)
        assert rs.dtype == 'timedelta64[ns]'

        df = DataFrame(dict(A=v1))
        td = Series([timedelta(days=i) for i in range(3)])
        assert td.dtype == 'timedelta64[ns]'

        # series on the rhs
        result = df['A'] - df['A'].shift()
        assert result.dtype == 'timedelta64[ns]'

        result = df['A'] + td
        assert result.dtype == 'M8[ns]'

        # scalar Timestamp on rhs
        maxa = df['A'].max()
        assert isinstance(maxa, Timestamp)

        resultb = df['A'] - df['A'].max()
        assert resultb.dtype == 'timedelta64[ns]'

        # timestamp on lhs
        result = resultb + df['A']
        values = [Timestamp('20111230'), Timestamp('20120101'),
                  Timestamp('20120103')]
        expected = Series(values, name='A')
        tm.assert_series_equal(result, expected)

        # datetimes on rhs
        result = df['A'] - datetime(2001, 1, 1)
        expected = Series(
            [timedelta(days=4017 + i) for i in range(3)], name='A')
        tm.assert_series_equal(result, expected)
        assert result.dtype == 'm8[ns]'

        d = datetime(2001, 1, 1, 3, 4)
        resulta = df['A'] - d
        assert resulta.dtype == 'm8[ns]'

        # roundtrip
        resultb = resulta + d
        tm.assert_series_equal(df['A'], resultb)

        # timedeltas on rhs
        td = timedelta(days=1)
        resulta = df['A'] + td
        resultb = resulta - td
        tm.assert_series_equal(resultb, df['A'])
        assert resultb.dtype == 'M8[ns]'

        # roundtrip
        td = timedelta(minutes=5, seconds=3)
        resulta = df['A'] + td
        resultb = resulta - td
        tm.assert_series_equal(df['A'], resultb)
        assert resultb.dtype == 'M8[ns]'

        # inplace
        value = rs[2] + np.timedelta64(timedelta(minutes=5, seconds=1))
        rs[2] += np.timedelta64(timedelta(minutes=5, seconds=1))
        assert rs[2] == value

    def test_timedelta64_ops_nat(self):
        # GH 11349
        timedelta_series = Series([NaT, Timedelta('1s')])
        nat_series_dtype_timedelta = Series([NaT, NaT],
                                            dtype='timedelta64[ns]')
        single_nat_dtype_timedelta = Series([NaT], dtype='timedelta64[ns]')

        # subtraction
        tm.assert_series_equal(timedelta_series - NaT,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(-NaT + timedelta_series,
                               nat_series_dtype_timedelta)

        tm.assert_series_equal(timedelta_series - single_nat_dtype_timedelta,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(-single_nat_dtype_timedelta + timedelta_series,
                               nat_series_dtype_timedelta)

        # addition
        tm.assert_series_equal(nat_series_dtype_timedelta + NaT,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(NaT + nat_series_dtype_timedelta,
                               nat_series_dtype_timedelta)

        tm.assert_series_equal(nat_series_dtype_timedelta +
                               single_nat_dtype_timedelta,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(single_nat_dtype_timedelta +
                               nat_series_dtype_timedelta,
                               nat_series_dtype_timedelta)

        tm.assert_series_equal(timedelta_series + NaT,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(NaT + timedelta_series,
                               nat_series_dtype_timedelta)

        tm.assert_series_equal(timedelta_series + single_nat_dtype_timedelta,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(single_nat_dtype_timedelta + timedelta_series,
                               nat_series_dtype_timedelta)

        tm.assert_series_equal(nat_series_dtype_timedelta + NaT,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(NaT + nat_series_dtype_timedelta,
                               nat_series_dtype_timedelta)

        tm.assert_series_equal(nat_series_dtype_timedelta +
                               single_nat_dtype_timedelta,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(single_nat_dtype_timedelta +
                               nat_series_dtype_timedelta,
                               nat_series_dtype_timedelta)

        # multiplication
        tm.assert_series_equal(nat_series_dtype_timedelta * 1.0,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(1.0 * nat_series_dtype_timedelta,
                               nat_series_dtype_timedelta)

        tm.assert_series_equal(timedelta_series * 1, timedelta_series)
        tm.assert_series_equal(1 * timedelta_series, timedelta_series)

        tm.assert_series_equal(timedelta_series * 1.5,
                               Series([NaT, Timedelta('1.5s')]))
        tm.assert_series_equal(1.5 * timedelta_series,
                               Series([NaT, Timedelta('1.5s')]))

        tm.assert_series_equal(timedelta_series * np.nan,
                               nat_series_dtype_timedelta)
        tm.assert_series_equal(np.nan * timedelta_series,
                               nat_series_dtype_timedelta)

        # division
        tm.assert_series_equal(timedelta_series / 2,
                               Series([NaT, Timedelta('0.5s')]))
        tm.assert_series_equal(timedelta_series / 2.0,
                               Series([NaT, Timedelta('0.5s')]))
        tm.assert_series_equal(timedelta_series / np.nan,
                               nat_series_dtype_timedelta)

    # -------------------------------------------------------------
    # Invalid Operations

    def test_td64arr_add_str_invalid(self, box_with_array):
        # GH#13624
        tdi = TimedeltaIndex(['1 day', '2 days'])
        tdi = tm.box_expected(tdi, box_with_array)

        with pytest.raises(TypeError):
            tdi + 'a'
        with pytest.raises(TypeError):
            'a' + tdi

    @pytest.mark.parametrize('other', [3.14, np.array([2.0, 3.0])])
    def test_td64arr_add_sub_float(self, box_with_array, other):
        tdi = TimedeltaIndex(['-1 days', '-1 days'])
        tdarr = tm.box_expected(tdi, box_with_array)

        with pytest.raises(TypeError):
            tdarr + other
        with pytest.raises(TypeError):
            other + tdarr
        with pytest.raises(TypeError):
            tdarr - other
        with pytest.raises(TypeError):
            other - tdarr

    @pytest.mark.parametrize('freq', [None, 'H'])
    def test_td64arr_sub_period(self, box_with_array, freq):
        # GH#13078
        # not supported, check TypeError
        p = pd.Period('2011-01-01', freq='D')
        idx = TimedeltaIndex(['1 hours', '2 hours'], freq=freq)
        idx = tm.box_expected(idx, box_with_array)

        with pytest.raises(TypeError):
            idx - p

        with pytest.raises(TypeError):
            p - idx

    @pytest.mark.parametrize('pi_freq', ['D', 'W', 'Q', 'H'])
    @pytest.mark.parametrize('tdi_freq', [None, 'H'])
    def test_td64arr_sub_pi(self, box_with_array, tdi_freq, pi_freq):
        # GH#20049 subtracting PeriodIndex should raise TypeError
        tdi = TimedeltaIndex(['1 hours', '2 hours'], freq=tdi_freq)
        dti = Timestamp('2018-03-07 17:16:40') + tdi
        pi = dti.to_period(pi_freq)

        # TODO: parametrize over box for pi?
        tdi = tm.box_expected(tdi, box_with_array)
        with pytest.raises(TypeError):
            tdi - pi

    # -------------------------------------------------------------
    # Binary operations td64 arraylike and datetime-like

    def test_td64arr_sub_timestamp_raises(self, box_with_array):
        idx = TimedeltaIndex(['1 day', '2 day'])
        idx = tm.box_expected(idx, box_with_array)

        msg = ("cannot subtract a datelike from|"
               "Could not operate|"
               "cannot perform operation")
        with pytest.raises(TypeError, match=msg):
            idx - Timestamp('2011-01-01')

    def test_td64arr_add_timestamp(self, box_with_array, tz_naive_fixture):
        # GH#23215

        # TODO: parametrize over scalar datetime types?
        tz = tz_naive_fixture
        other = Timestamp('2011-01-01', tz=tz)

        idx = TimedeltaIndex(['1 day', '2 day'])
        expected = DatetimeIndex(['2011-01-02', '2011-01-03'], tz=tz)

        # FIXME: fails with transpose=True because of tz-aware DataFrame
        #  transpose bug
        idx = tm.box_expected(idx, box_with_array, transpose=False)
        expected = tm.box_expected(expected, box_with_array, transpose=False)

        result = idx + other
        tm.assert_equal(result, expected)

        result = other + idx
        tm.assert_equal(result, expected)

    def test_td64arr_add_sub_timestamp(self, box_with_array):
        # GH#11925
        ts = Timestamp('2012-01-01')
        # TODO: parametrize over types of datetime scalar?

        tdi = timedelta_range('1 day', periods=3)
        expected = pd.date_range('2012-01-02', periods=3)

        tdarr = tm.box_expected(tdi, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        tm.assert_equal(ts + tdarr, expected)
        tm.assert_equal(tdarr + ts, expected)

        expected2 = pd.date_range('2011-12-31', periods=3, freq='-1D')
        expected2 = tm.box_expected(expected2, box_with_array)

        tm.assert_equal(ts - tdarr, expected2)
        tm.assert_equal(ts + (-tdarr), expected2)

        with pytest.raises(TypeError):
            tdarr - ts

    def test_tdi_sub_dt64_array(self, box_with_array):
        dti = pd.date_range('2016-01-01', periods=3)
        tdi = dti - dti.shift(1)
        dtarr = dti.values
        expected = pd.DatetimeIndex(dtarr) - tdi

        tdi = tm.box_expected(tdi, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        with pytest.raises(TypeError):
            tdi - dtarr

        # TimedeltaIndex.__rsub__
        result = dtarr - tdi
        tm.assert_equal(result, expected)

    def test_tdi_add_dt64_array(self, box_with_array):
        dti = pd.date_range('2016-01-01', periods=3)
        tdi = dti - dti.shift(1)
        dtarr = dti.values
        expected = pd.DatetimeIndex(dtarr) + tdi

        tdi = tm.box_expected(tdi, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        result = tdi + dtarr
        tm.assert_equal(result, expected)
        result = dtarr + tdi
        tm.assert_equal(result, expected)

    def test_td64arr_add_datetime64_nat(self, box_with_array):
        # GH#23215
        other = np.datetime64('NaT')

        tdi = timedelta_range('1 day', periods=3)
        expected = pd.DatetimeIndex(["NaT", "NaT", "NaT"])

        tdser = tm.box_expected(tdi, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        tm.assert_equal(tdser + other, expected)
        tm.assert_equal(other + tdser, expected)

    # ------------------------------------------------------------------
    # Operations with int-like others

    def test_td64arr_add_int_series_invalid(self, box):
        tdser = pd.Series(['59 Days', '59 Days', 'NaT'], dtype='m8[ns]')
        tdser = tm.box_expected(tdser, box)
        err = TypeError if box is not pd.Index else NullFrequencyError
        int_ser = Series([2, 3, 4])

        with pytest.raises(err):
            tdser + int_ser
        with pytest.raises(err):
            int_ser + tdser
        with pytest.raises(err):
            tdser - int_ser
        with pytest.raises(err):
            int_ser - tdser

    def test_td64arr_add_intlike(self, box_with_array):
        # GH#19123
        tdi = TimedeltaIndex(['59 days', '59 days', 'NaT'])
        ser = tm.box_expected(tdi, box_with_array)

        err = TypeError
        if box_with_array in [pd.Index, tm.to_array]:
            err = NullFrequencyError

        other = Series([20, 30, 40], dtype='uint8')

        # TODO: separate/parametrize
        with pytest.raises(err):
            ser + 1
        with pytest.raises(err):
            ser - 1

        with pytest.raises(err):
            ser + other
        with pytest.raises(err):
            ser - other

        with pytest.raises(err):
            ser + np.array(other)
        with pytest.raises(err):
            ser - np.array(other)

        with pytest.raises(err):
            ser + pd.Index(other)
        with pytest.raises(err):
            ser - pd.Index(other)

    @pytest.mark.parametrize('scalar', [1, 1.5, np.array(2)])
    def test_td64arr_add_sub_numeric_scalar_invalid(self, box_with_array,
                                                    scalar):
        box = box_with_array

        tdser = pd.Series(['59 Days', '59 Days', 'NaT'], dtype='m8[ns]')
        tdser = tm.box_expected(tdser, box)
        err = TypeError
        if box in [pd.Index, tm.to_array] and not isinstance(scalar, float):
            err = NullFrequencyError

        with pytest.raises(err):
            tdser + scalar
        with pytest.raises(err):
            scalar + tdser
        with pytest.raises(err):
            tdser - scalar
        with pytest.raises(err):
            scalar - tdser

    @pytest.mark.parametrize('dtype', ['int64', 'int32', 'int16',
                                       'uint64', 'uint32', 'uint16', 'uint8',
                                       'float64', 'float32', 'float16'])
    @pytest.mark.parametrize('vec', [
        np.array([1, 2, 3]),
        pd.Index([1, 2, 3]),
        Series([1, 2, 3])
        # TODO: Add DataFrame in here?
    ], ids=lambda x: type(x).__name__)
    def test_td64arr_add_sub_numeric_arr_invalid(self, box, vec, dtype):
        tdser = pd.Series(['59 Days', '59 Days', 'NaT'], dtype='m8[ns]')
        tdser = tm.box_expected(tdser, box)
        err = TypeError
        if box is pd.Index and not dtype.startswith('float'):
            err = NullFrequencyError

        vector = vec.astype(dtype)
        with pytest.raises(err):
            tdser + vector
        with pytest.raises(err):
            vector + tdser
        with pytest.raises(err):
            tdser - vector
        with pytest.raises(err):
            vector - tdser

    # ------------------------------------------------------------------
    # Operations with timedelta-like others

    # TODO: this was taken from tests.series.test_ops; de-duplicate
    @pytest.mark.parametrize('scalar_td', [timedelta(minutes=5, seconds=4),
                                           Timedelta(minutes=5, seconds=4),
                                           Timedelta('5m4s').to_timedelta64()])
    def test_operators_timedelta64_with_timedelta(self, scalar_td):
        # smoke tests
        td1 = Series([timedelta(minutes=5, seconds=3)] * 3)
        td1.iloc[2] = np.nan

        td1 + scalar_td
        scalar_td + td1
        td1 - scalar_td
        scalar_td - td1
        td1 / scalar_td
        scalar_td / td1

    # TODO: this was taken from tests.series.test_ops; de-duplicate
    def test_timedelta64_operations_with_timedeltas(self):
        # td operate with td
        td1 = Series([timedelta(minutes=5, seconds=3)] * 3)
        td2 = timedelta(minutes=5, seconds=4)
        result = td1 - td2
        expected = (Series([timedelta(seconds=0)] * 3) -
                    Series([timedelta(seconds=1)] * 3))
        assert result.dtype == 'm8[ns]'
        tm.assert_series_equal(result, expected)

        result2 = td2 - td1
        expected = (Series([timedelta(seconds=1)] * 3) -
                    Series([timedelta(seconds=0)] * 3))
        tm.assert_series_equal(result2, expected)

        # roundtrip
        tm.assert_series_equal(result + td2, td1)

        # Now again, using pd.to_timedelta, which should build
        # a Series or a scalar, depending on input.
        td1 = Series(pd.to_timedelta(['00:05:03'] * 3))
        td2 = pd.to_timedelta('00:05:04')
        result = td1 - td2
        expected = (Series([timedelta(seconds=0)] * 3) -
                    Series([timedelta(seconds=1)] * 3))
        assert result.dtype == 'm8[ns]'
        tm.assert_series_equal(result, expected)

        result2 = td2 - td1
        expected = (Series([timedelta(seconds=1)] * 3) -
                    Series([timedelta(seconds=0)] * 3))
        tm.assert_series_equal(result2, expected)

        # roundtrip
        tm.assert_series_equal(result + td2, td1)

    def test_td64arr_add_td64_array(self, box):
        dti = pd.date_range('2016-01-01', periods=3)
        tdi = dti - dti.shift(1)
        tdarr = tdi.values

        expected = 2 * tdi
        tdi = tm.box_expected(tdi, box)
        expected = tm.box_expected(expected, box)

        result = tdi + tdarr
        tm.assert_equal(result, expected)
        result = tdarr + tdi
        tm.assert_equal(result, expected)

    def test_td64arr_sub_td64_array(self, box):
        dti = pd.date_range('2016-01-01', periods=3)
        tdi = dti - dti.shift(1)
        tdarr = tdi.values

        expected = 0 * tdi
        tdi = tm.box_expected(tdi, box)
        expected = tm.box_expected(expected, box)

        result = tdi - tdarr
        tm.assert_equal(result, expected)
        result = tdarr - tdi
        tm.assert_equal(result, expected)

    # TODO: parametrize over [add, sub, radd, rsub]?
    @pytest.mark.parametrize('names', [(None, None, None),
                                       ('Egon', 'Venkman', None),
                                       ('NCC1701D', 'NCC1701D', 'NCC1701D')])
    def test_td64arr_add_sub_tdi(self, box, names):
        # GH#17250 make sure result dtype is correct
        # GH#19043 make sure names are propagated correctly
        if box is pd.DataFrame and names[1] == 'Venkman':
            pytest.skip("Name propagation for DataFrame does not behave like "
                        "it does for Index/Series")

        tdi = TimedeltaIndex(['0 days', '1 day'], name=names[0])
        ser = Series([Timedelta(hours=3), Timedelta(hours=4)], name=names[1])
        expected = Series([Timedelta(hours=3), Timedelta(days=1, hours=4)],
                          name=names[2])

        ser = tm.box_expected(ser, box)
        expected = tm.box_expected(expected, box)

        result = tdi + ser
        tm.assert_equal(result, expected)
        if box is not pd.DataFrame:
            assert result.dtype == 'timedelta64[ns]'
        else:
            assert result.dtypes[0] == 'timedelta64[ns]'

        result = ser + tdi
        tm.assert_equal(result, expected)
        if box is not pd.DataFrame:
            assert result.dtype == 'timedelta64[ns]'
        else:
            assert result.dtypes[0] == 'timedelta64[ns]'

        expected = Series([Timedelta(hours=-3), Timedelta(days=1, hours=-4)],
                          name=names[2])
        expected = tm.box_expected(expected, box)

        result = tdi - ser
        tm.assert_equal(result, expected)
        if box is not pd.DataFrame:
            assert result.dtype == 'timedelta64[ns]'
        else:
            assert result.dtypes[0] == 'timedelta64[ns]'

        result = ser - tdi
        tm.assert_equal(result, -expected)
        if box is not pd.DataFrame:
            assert result.dtype == 'timedelta64[ns]'
        else:
            assert result.dtypes[0] == 'timedelta64[ns]'

    def test_td64arr_add_sub_td64_nat(self, box):
        # GH#23320 special handling for timedelta64("NaT")
        tdi = pd.TimedeltaIndex([NaT, Timedelta('1s')])
        other = np.timedelta64("NaT")
        expected = pd.TimedeltaIndex(["NaT"] * 2)

        obj = tm.box_expected(tdi, box)
        expected = tm.box_expected(expected, box)

        result = obj + other
        tm.assert_equal(result, expected)
        result = other + obj
        tm.assert_equal(result, expected)
        result = obj - other
        tm.assert_equal(result, expected)
        result = other - obj
        tm.assert_equal(result, expected)

    def test_td64arr_sub_NaT(self, box):
        # GH#18808
        ser = Series([NaT, Timedelta('1s')])
        expected = Series([NaT, NaT], dtype='timedelta64[ns]')

        ser = tm.box_expected(ser, box)
        expected = tm.box_expected(expected, box)

        res = ser - pd.NaT
        tm.assert_equal(res, expected)

    def test_td64arr_add_timedeltalike(self, two_hours, box):
        # only test adding/sub offsets as + is now numeric
        rng = timedelta_range('1 days', '10 days')
        expected = timedelta_range('1 days 02:00:00', '10 days 02:00:00',
                                   freq='D')
        rng = tm.box_expected(rng, box)
        expected = tm.box_expected(expected, box)

        result = rng + two_hours
        tm.assert_equal(result, expected)

    def test_td64arr_sub_timedeltalike(self, two_hours, box):
        # only test adding/sub offsets as - is now numeric
        rng = timedelta_range('1 days', '10 days')
        expected = timedelta_range('0 days 22:00:00', '9 days 22:00:00')

        rng = tm.box_expected(rng, box)
        expected = tm.box_expected(expected, box)

        result = rng - two_hours
        tm.assert_equal(result, expected)

    # ------------------------------------------------------------------
    # __add__/__sub__ with DateOffsets and arrays of DateOffsets

    # TODO: this was taken from tests.series.test_operators; de-duplicate
    def test_timedelta64_operations_with_DateOffset(self):
        # GH#10699
        td = Series([timedelta(minutes=5, seconds=3)] * 3)
        result = td + pd.offsets.Minute(1)
        expected = Series([timedelta(minutes=6, seconds=3)] * 3)
        tm.assert_series_equal(result, expected)

        result = td - pd.offsets.Minute(1)
        expected = Series([timedelta(minutes=4, seconds=3)] * 3)
        tm.assert_series_equal(result, expected)

        with tm.assert_produces_warning(PerformanceWarning):
            result = td + Series([pd.offsets.Minute(1), pd.offsets.Second(3),
                                  pd.offsets.Hour(2)])
        expected = Series([timedelta(minutes=6, seconds=3),
                           timedelta(minutes=5, seconds=6),
                           timedelta(hours=2, minutes=5, seconds=3)])
        tm.assert_series_equal(result, expected)

        result = td + pd.offsets.Minute(1) + pd.offsets.Second(12)
        expected = Series([timedelta(minutes=6, seconds=15)] * 3)
        tm.assert_series_equal(result, expected)

        # valid DateOffsets
        for do in ['Hour', 'Minute', 'Second', 'Day', 'Micro', 'Milli',
                   'Nano']:
            op = getattr(pd.offsets, do)
            td + op(5)
            op(5) + td
            td - op(5)
            op(5) - td

    @pytest.mark.parametrize('names', [(None, None, None),
                                       ('foo', 'bar', None),
                                       ('foo', 'foo', 'foo')])
    def test_td64arr_add_offset_index(self, names, box):
        # GH#18849, GH#19744
        if box is pd.DataFrame and names[1] == 'bar':
            pytest.skip("Name propagation for DataFrame does not behave like "
                        "it does for Index/Series")

        tdi = TimedeltaIndex(['1 days 00:00:00', '3 days 04:00:00'],
                             name=names[0])
        other = pd.Index([pd.offsets.Hour(n=1), pd.offsets.Minute(n=-2)],
                         name=names[1])

        expected = TimedeltaIndex([tdi[n] + other[n] for n in range(len(tdi))],
                                  freq='infer', name=names[2])
        tdi = tm.box_expected(tdi, box)
        expected = tm.box_expected(expected, box)

        # The DataFrame operation is transposed and so operates as separate
        #  scalar operations, which do not issue a PerformanceWarning
        warn = PerformanceWarning if box is not pd.DataFrame else None
        with tm.assert_produces_warning(warn):
            res = tdi + other
        tm.assert_equal(res, expected)

        with tm.assert_produces_warning(warn):
            res2 = other + tdi
        tm.assert_equal(res2, expected)

    # TODO: combine with test_td64arr_add_offset_index by parametrizing
    # over second box?
    def test_td64arr_add_offset_array(self, box):
        # GH#18849
        tdi = TimedeltaIndex(['1 days 00:00:00', '3 days 04:00:00'])
        other = np.array([pd.offsets.Hour(n=1), pd.offsets.Minute(n=-2)])

        expected = TimedeltaIndex([tdi[n] + other[n] for n in range(len(tdi))],
                                  freq='infer')

        tdi = tm.box_expected(tdi, box)
        expected = tm.box_expected(expected, box)

        # The DataFrame operation is transposed and so operates as separate
        #  scalar operations, which do not issue a PerformanceWarning
        warn = PerformanceWarning if box is not pd.DataFrame else None
        with tm.assert_produces_warning(warn):
            res = tdi + other
        tm.assert_equal(res, expected)

        with tm.assert_produces_warning(warn):
            res2 = other + tdi
        tm.assert_equal(res2, expected)

    @pytest.mark.parametrize('names', [(None, None, None),
                                       ('foo', 'bar', None),
                                       ('foo', 'foo', 'foo')])
    def test_td64arr_sub_offset_index(self, names, box):
        # GH#18824, GH#19744
        if box is pd.DataFrame and names[1] == 'bar':
            pytest.skip("Name propagation for DataFrame does not behave like "
                        "it does for Index/Series")

        tdi = TimedeltaIndex(['1 days 00:00:00', '3 days 04:00:00'],
                             name=names[0])
        other = pd.Index([pd.offsets.Hour(n=1), pd.offsets.Minute(n=-2)],
                         name=names[1])

        expected = TimedeltaIndex([tdi[n] - other[n] for n in range(len(tdi))],
                                  freq='infer', name=names[2])

        tdi = tm.box_expected(tdi, box)
        expected = tm.box_expected(expected, box)

        # The DataFrame operation is transposed and so operates as separate
        #  scalar operations, which do not issue a PerformanceWarning
        warn = PerformanceWarning if box is not pd.DataFrame else None
        with tm.assert_produces_warning(warn):
            res = tdi - other
        tm.assert_equal(res, expected)

    def test_td64arr_sub_offset_array(self, box_with_array):
        # GH#18824
        tdi = TimedeltaIndex(['1 days 00:00:00', '3 days 04:00:00'])
        other = np.array([pd.offsets.Hour(n=1), pd.offsets.Minute(n=-2)])

        expected = TimedeltaIndex([tdi[n] - other[n] for n in range(len(tdi))],
                                  freq='infer')

        tdi = tm.box_expected(tdi, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        # The DataFrame operation is transposed and so operates as separate
        #  scalar operations, which do not issue a PerformanceWarning
        warn = None if box_with_array is pd.DataFrame else PerformanceWarning
        with tm.assert_produces_warning(warn):
            res = tdi - other
        tm.assert_equal(res, expected)

    @pytest.mark.parametrize('names', [(None, None, None),
                                       ('foo', 'bar', None),
                                       ('foo', 'foo', 'foo')])
    def test_td64arr_with_offset_series(self, names, box_df_fail):
        # GH#18849
        box = box_df_fail
        box2 = Series if box in [pd.Index, tm.to_array] else box

        tdi = TimedeltaIndex(['1 days 00:00:00', '3 days 04:00:00'],
                             name=names[0])
        other = Series([pd.offsets.Hour(n=1), pd.offsets.Minute(n=-2)],
                       name=names[1])

        expected_add = Series([tdi[n] + other[n] for n in range(len(tdi))],
                              name=names[2])
        tdi = tm.box_expected(tdi, box)
        expected_add = tm.box_expected(expected_add, box2)

        with tm.assert_produces_warning(PerformanceWarning):
            res = tdi + other
        tm.assert_equal(res, expected_add)

        with tm.assert_produces_warning(PerformanceWarning):
            res2 = other + tdi
        tm.assert_equal(res2, expected_add)

        # TODO: separate/parametrize add/sub test?
        expected_sub = Series([tdi[n] - other[n] for n in range(len(tdi))],
                              name=names[2])
        expected_sub = tm.box_expected(expected_sub, box2)

        with tm.assert_produces_warning(PerformanceWarning):
            res3 = tdi - other
        tm.assert_equal(res3, expected_sub)

    @pytest.mark.parametrize('obox', [np.array, pd.Index, pd.Series])
    def test_td64arr_addsub_anchored_offset_arraylike(self, obox,
                                                      box_with_array):
        # GH#18824
        tdi = TimedeltaIndex(['1 days 00:00:00', '3 days 04:00:00'])
        tdi = tm.box_expected(tdi, box_with_array)

        anchored = obox([pd.offsets.MonthEnd(), pd.offsets.Day(n=2)])

        # addition/subtraction ops with anchored offsets should issue
        # a PerformanceWarning and _then_ raise a TypeError.
        with pytest.raises(TypeError):
            with tm.assert_produces_warning(PerformanceWarning):
                tdi + anchored
        with pytest.raises(TypeError):
            with tm.assert_produces_warning(PerformanceWarning):
                anchored + tdi
        with pytest.raises(TypeError):
            with tm.assert_produces_warning(PerformanceWarning):
                tdi - anchored
        with pytest.raises(TypeError):
            with tm.assert_produces_warning(PerformanceWarning):
                anchored - tdi


class TestTimedeltaArraylikeMulDivOps(object):
    # Tests for timedelta64[ns]
    # __mul__, __rmul__, __div__, __rdiv__, __floordiv__, __rfloordiv__

    # TODO: Moved from tests.series.test_operators; needs cleanup
    @pytest.mark.parametrize("m", [1, 3, 10])
    @pytest.mark.parametrize("unit", ['D', 'h', 'm', 's', 'ms', 'us', 'ns'])
    def test_timedelta64_conversions(self, m, unit):
        startdate = Series(pd.date_range('2013-01-01', '2013-01-03'))
        enddate = Series(pd.date_range('2013-03-01', '2013-03-03'))

        ser = enddate - startdate
        ser[2] = np.nan

        # op
        expected = Series([x / np.timedelta64(m, unit) for x in ser])
        result = ser / np.timedelta64(m, unit)
        tm.assert_series_equal(result, expected)

        # reverse op
        expected = Series([Timedelta(np.timedelta64(m, unit)) / x
                           for x in ser])
        result = np.timedelta64(m, unit) / ser
        tm.assert_series_equal(result, expected)

    # ------------------------------------------------------------------
    # Multiplication
    # organized with scalar others first, then array-like

    def test_td64arr_mul_int(self, box_with_array):
        idx = TimedeltaIndex(np.arange(5, dtype='int64'))
        idx = tm.box_expected(idx, box_with_array)

        result = idx * 1
        tm.assert_equal(result, idx)

        result = 1 * idx
        tm.assert_equal(result, idx)

    def test_td64arr_mul_tdlike_scalar_raises(self, two_hours, box_with_array):
        rng = timedelta_range('1 days', '10 days', name='foo')
        rng = tm.box_expected(rng, box_with_array)
        with pytest.raises(TypeError):
            rng * two_hours

    def test_tdi_mul_int_array_zerodim(self, box_with_array):
        rng5 = np.arange(5, dtype='int64')
        idx = TimedeltaIndex(rng5)
        expected = TimedeltaIndex(rng5 * 5)

        idx = tm.box_expected(idx, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        result = idx * np.array(5, dtype='int64')
        tm.assert_equal(result, expected)

    def test_tdi_mul_int_array(self, box_with_array):
        rng5 = np.arange(5, dtype='int64')
        idx = TimedeltaIndex(rng5)
        expected = TimedeltaIndex(rng5 ** 2)

        idx = tm.box_expected(idx, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        result = idx * rng5
        tm.assert_equal(result, expected)

    def test_tdi_mul_int_series(self, box_with_array):
        box = box_with_array
        xbox = pd.Series if box in [pd.Index, tm.to_array] else box

        idx = TimedeltaIndex(np.arange(5, dtype='int64'))
        expected = TimedeltaIndex(np.arange(5, dtype='int64') ** 2)

        idx = tm.box_expected(idx, box)
        expected = tm.box_expected(expected, xbox)

        result = idx * pd.Series(np.arange(5, dtype='int64'))
        tm.assert_equal(result, expected)

    def test_tdi_mul_float_series(self, box_with_array):
        box = box_with_array
        xbox = pd.Series if box in [pd.Index, tm.to_array] else box

        idx = TimedeltaIndex(np.arange(5, dtype='int64'))
        idx = tm.box_expected(idx, box)

        rng5f = np.arange(5, dtype='float64')
        expected = TimedeltaIndex(rng5f * (rng5f + 1.0))
        expected = tm.box_expected(expected, xbox)

        result = idx * Series(rng5f + 1.0)
        tm.assert_equal(result, expected)

    # TODO: Put Series/DataFrame in others?
    @pytest.mark.parametrize('other', [
        np.arange(1, 11),
        pd.Int64Index(range(1, 11)),
        pd.UInt64Index(range(1, 11)),
        pd.Float64Index(range(1, 11)),
        pd.RangeIndex(1, 11)
    ], ids=lambda x: type(x).__name__)
    def test_tdi_rmul_arraylike(self, other, box_with_array):
        box = box_with_array
        xbox = get_upcast_box(box, other)

        tdi = TimedeltaIndex(['1 Day'] * 10)
        expected = timedelta_range('1 days', '10 days')
        expected._data.freq = None

        tdi = tm.box_expected(tdi, box)
        expected = tm.box_expected(expected, xbox)

        result = other * tdi
        tm.assert_equal(result, expected)
        commute = tdi * other
        tm.assert_equal(commute, expected)

    # ------------------------------------------------------------------
    # __div__, __rdiv__

    def test_td64arr_div_nat_invalid(self, box_with_array):
        # don't allow division by NaT (maybe could in the future)
        rng = timedelta_range('1 days', '10 days', name='foo')
        rng = tm.box_expected(rng, box_with_array)

        with pytest.raises(TypeError, match='true_divide cannot use operands'):
            rng / pd.NaT
        with pytest.raises(TypeError, match='Cannot divide NaTType by'):
            pd.NaT / rng

    def test_td64arr_div_td64nat(self, box_with_array):
        # GH#23829
        rng = timedelta_range('1 days', '10 days',)
        rng = tm.box_expected(rng, box_with_array)

        other = np.timedelta64('NaT')

        expected = np.array([np.nan] * 10)
        expected = tm.box_expected(expected, box_with_array)

        result = rng / other
        tm.assert_equal(result, expected)

        result = other / rng
        tm.assert_equal(result, expected)

    def test_td64arr_div_int(self, box_with_array):
        idx = TimedeltaIndex(np.arange(5, dtype='int64'))
        idx = tm.box_expected(idx, box_with_array)

        result = idx / 1
        tm.assert_equal(result, idx)

        with pytest.raises(TypeError, match='Cannot divide'):
            # GH#23829
            1 / idx

    def test_td64arr_div_tdlike_scalar(self, two_hours, box_with_array):
        # GH#20088, GH#22163 ensure DataFrame returns correct dtype
        rng = timedelta_range('1 days', '10 days', name='foo')
        expected = pd.Float64Index((np.arange(10) + 1) * 12, name='foo')

        rng = tm.box_expected(rng, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        result = rng / two_hours
        tm.assert_equal(result, expected)

        result = two_hours / rng
        expected = 1 / expected
        tm.assert_equal(result, expected)

    def test_td64arr_div_tdlike_scalar_with_nat(self, two_hours,
                                                box_with_array):
        rng = TimedeltaIndex(['1 days', pd.NaT, '2 days'], name='foo')
        expected = pd.Float64Index([12, np.nan, 24], name='foo')

        rng = tm.box_expected(rng, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        result = rng / two_hours
        tm.assert_equal(result, expected)

        result = two_hours / rng
        expected = 1 / expected
        tm.assert_equal(result, expected)

    def test_td64arr_div_td64_ndarray(self, box_with_array):
        # GH#22631
        rng = TimedeltaIndex(['1 days', pd.NaT, '2 days'])
        expected = pd.Float64Index([12, np.nan, 24])

        rng = tm.box_expected(rng, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        other = np.array([2, 4, 2], dtype='m8[h]')
        result = rng / other
        tm.assert_equal(result, expected)

        result = rng / tm.box_expected(other, box_with_array)
        tm.assert_equal(result, expected)

        result = rng / other.astype(object)
        tm.assert_equal(result, expected)

        result = rng / list(other)
        tm.assert_equal(result, expected)

        # reversed op
        expected = 1 / expected
        result = other / rng
        tm.assert_equal(result, expected)

        result = tm.box_expected(other, box_with_array) / rng
        tm.assert_equal(result, expected)

        result = other.astype(object) / rng
        tm.assert_equal(result, expected)

        result = list(other) / rng
        tm.assert_equal(result, expected)

    def test_tdarr_div_length_mismatch(self, box_with_array):
        rng = TimedeltaIndex(['1 days', pd.NaT, '2 days'])
        mismatched = [1, 2, 3, 4]

        rng = tm.box_expected(rng, box_with_array)
        for obj in [mismatched, mismatched[:2]]:
            # one shorter, one longer
            for other in [obj, np.array(obj), pd.Index(obj)]:
                with pytest.raises(ValueError):
                    rng / other
                with pytest.raises(ValueError):
                    other / rng

    # ------------------------------------------------------------------
    # __floordiv__, __rfloordiv__

    def test_td64arr_floordiv_tdscalar(self, box_with_array, scalar_td):
        # GH#18831
        td1 = Series([timedelta(minutes=5, seconds=3)] * 3)
        td1.iloc[2] = np.nan

        expected = Series([0, 0, np.nan])

        td1 = tm.box_expected(td1, box_with_array, transpose=False)
        expected = tm.box_expected(expected, box_with_array, transpose=False)

        result = td1 // scalar_td
        tm.assert_equal(result, expected)

    def test_td64arr_rfloordiv_tdscalar(self, box_with_array, scalar_td):
        # GH#18831
        td1 = Series([timedelta(minutes=5, seconds=3)] * 3)
        td1.iloc[2] = np.nan

        expected = Series([1, 1, np.nan])

        td1 = tm.box_expected(td1, box_with_array, transpose=False)
        expected = tm.box_expected(expected, box_with_array, transpose=False)

        result = scalar_td // td1
        tm.assert_equal(result, expected)

    def test_td64arr_rfloordiv_tdscalar_explicit(self, box_with_array,
                                                 scalar_td):
        # GH#18831
        td1 = Series([timedelta(minutes=5, seconds=3)] * 3)
        td1.iloc[2] = np.nan

        expected = Series([1, 1, np.nan])

        td1 = tm.box_expected(td1, box_with_array, transpose=False)
        expected = tm.box_expected(expected, box_with_array, transpose=False)

        # We can test __rfloordiv__ using this syntax,
        # see `test_timedelta_rfloordiv`
        result = td1.__rfloordiv__(scalar_td)
        tm.assert_equal(result, expected)

    def test_td64arr_floordiv_int(self, box_with_array):
        idx = TimedeltaIndex(np.arange(5, dtype='int64'))
        idx = tm.box_expected(idx, box_with_array)
        result = idx // 1
        tm.assert_equal(result, idx)

        pattern = ('floor_divide cannot use operands|'
                   'Cannot divide int by Timedelta*')
        with pytest.raises(TypeError, match=pattern):
            1 // idx

    def test_td64arr_floordiv_tdlike_scalar(self, two_hours, box_with_array):
        tdi = timedelta_range('1 days', '10 days', name='foo')
        expected = pd.Int64Index((np.arange(10) + 1) * 12, name='foo')

        tdi = tm.box_expected(tdi, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        result = tdi // two_hours
        tm.assert_equal(result, expected)

    # TODO: Is this redundant with test_td64arr_floordiv_tdlike_scalar?
    @pytest.mark.parametrize('scalar_td', [
        timedelta(minutes=10, seconds=7),
        Timedelta('10m7s'),
        Timedelta('10m7s').to_timedelta64()
    ], ids=lambda x: type(x).__name__)
    def test_td64arr_rfloordiv_tdlike_scalar(self, scalar_td, box_with_array):
        # GH#19125
        tdi = TimedeltaIndex(['00:05:03', '00:05:03', pd.NaT], freq=None)
        expected = pd.Index([2.0, 2.0, np.nan])

        tdi = tm.box_expected(tdi, box_with_array, transpose=False)
        expected = tm.box_expected(expected, box_with_array, transpose=False)

        res = tdi.__rfloordiv__(scalar_td)
        tm.assert_equal(res, expected)

        expected = pd.Index([0.0, 0.0, np.nan])
        expected = tm.box_expected(expected, box_with_array, transpose=False)

        res = tdi // (scalar_td)
        tm.assert_equal(res, expected)

    # ------------------------------------------------------------------
    # mod, divmod
    # TODO: operations with timedelta-like arrays, numeric arrays,
    #  reversed ops

    def test_td64arr_mod_tdscalar(self, box_with_array, three_days):
        tdi = timedelta_range('1 Day', '9 days')
        tdarr = tm.box_expected(tdi, box_with_array)

        expected = TimedeltaIndex(['1 Day', '2 Days', '0 Days'] * 3)
        expected = tm.box_expected(expected, box_with_array)

        result = tdarr % three_days
        tm.assert_equal(result, expected)

        if box_with_array is pd.DataFrame:
            pytest.xfail("DataFrame does not have __divmod__ or __rdivmod__")

        result = divmod(tdarr, three_days)
        tm.assert_equal(result[1], expected)
        tm.assert_equal(result[0], tdarr // three_days)

    def test_td64arr_mod_int(self, box_with_array):
        tdi = timedelta_range('1 ns', '10 ns', periods=10)
        tdarr = tm.box_expected(tdi, box_with_array)

        expected = TimedeltaIndex(['1 ns', '0 ns'] * 5)
        expected = tm.box_expected(expected, box_with_array)

        result = tdarr % 2
        tm.assert_equal(result, expected)

        with pytest.raises(TypeError):
            2 % tdarr

        if box_with_array is pd.DataFrame:
            pytest.xfail("DataFrame does not have __divmod__ or __rdivmod__")

        result = divmod(tdarr, 2)
        tm.assert_equal(result[1], expected)
        tm.assert_equal(result[0], tdarr // 2)

    def test_td64arr_rmod_tdscalar(self, box_with_array, three_days):
        tdi = timedelta_range('1 Day', '9 days')
        tdarr = tm.box_expected(tdi, box_with_array)

        expected = ['0 Days', '1 Day', '0 Days'] + ['3 Days'] * 6
        expected = TimedeltaIndex(expected)
        expected = tm.box_expected(expected, box_with_array)

        result = three_days % tdarr
        tm.assert_equal(result, expected)

        if box_with_array is pd.DataFrame:
            pytest.xfail("DataFrame does not have __divmod__ or __rdivmod__")

        result = divmod(three_days, tdarr)
        tm.assert_equal(result[1], expected)
        tm.assert_equal(result[0], three_days // tdarr)

    # ------------------------------------------------------------------
    # Operations with invalid others

    def test_td64arr_mul_tdscalar_invalid(self, box_with_array, scalar_td):
        td1 = Series([timedelta(minutes=5, seconds=3)] * 3)
        td1.iloc[2] = np.nan

        td1 = tm.box_expected(td1, box_with_array)

        # check that we are getting a TypeError
        # with 'operate' (from core/ops.py) for the ops that are not
        # defined
        pattern = 'operate|unsupported|cannot|not supported'
        with pytest.raises(TypeError, match=pattern):
            td1 * scalar_td
        with pytest.raises(TypeError, match=pattern):
            scalar_td * td1

    def test_td64arr_mul_too_short_raises(self, box_with_array):
        idx = TimedeltaIndex(np.arange(5, dtype='int64'))
        idx = tm.box_expected(idx, box_with_array)
        with pytest.raises(TypeError):
            idx * idx[:3]
        with pytest.raises(ValueError):
            idx * np.array([1, 2])

    def test_td64arr_mul_td64arr_raises(self, box_with_array):
        idx = TimedeltaIndex(np.arange(5, dtype='int64'))
        idx = tm.box_expected(idx, box_with_array)
        with pytest.raises(TypeError):
            idx * idx

    # ------------------------------------------------------------------
    # Operations with numeric others

    @pytest.mark.parametrize('one', [1, np.array(1), 1.0, np.array(1.0)])
    def test_td64arr_mul_numeric_scalar(self, box_with_array, one):
        # GH#4521
        # divide/multiply by integers
        tdser = pd.Series(['59 Days', '59 Days', 'NaT'], dtype='m8[ns]')
        expected = Series(['-59 Days', '-59 Days', 'NaT'],
                          dtype='timedelta64[ns]')

        tdser = tm.box_expected(tdser, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        result = tdser * (-one)
        tm.assert_equal(result, expected)
        result = (-one) * tdser
        tm.assert_equal(result, expected)

        expected = Series(['118 Days', '118 Days', 'NaT'],
                          dtype='timedelta64[ns]')
        expected = tm.box_expected(expected, box_with_array)

        result = tdser * (2 * one)
        tm.assert_equal(result, expected)
        result = (2 * one) * tdser
        tm.assert_equal(result, expected)

    @pytest.mark.parametrize('two', [2, 2.0, np.array(2), np.array(2.0)])
    def test_td64arr_div_numeric_scalar(self, box_with_array, two):
        # GH#4521
        # divide/multiply by integers
        tdser = pd.Series(['59 Days', '59 Days', 'NaT'], dtype='m8[ns]')
        expected = Series(['29.5D', '29.5D', 'NaT'], dtype='timedelta64[ns]')

        tdser = tm.box_expected(tdser, box_with_array)
        expected = tm.box_expected(expected, box_with_array)

        result = tdser / two
        tm.assert_equal(result, expected)

        with pytest.raises(TypeError, match='Cannot divide'):
            two / tdser

    @pytest.mark.parametrize('dtype', ['int64', 'int32', 'int16',
                                       'uint64', 'uint32', 'uint16', 'uint8',
                                       'float64', 'float32', 'float16'])
    @pytest.mark.parametrize('vector', [np.array([20, 30, 40]),
                                        pd.Index([20, 30, 40]),
                                        Series([20, 30, 40])],
                             ids=lambda x: type(x).__name__)
    def test_td64arr_rmul_numeric_array(self, box_with_array, vector, dtype):
        # GH#4521
        # divide/multiply by integers
        xbox = get_upcast_box(box_with_array, vector)

        tdser = pd.Series(['59 Days', '59 Days', 'NaT'], dtype='m8[ns]')
        vector = vector.astype(dtype)

        expected = Series(['1180 Days', '1770 Days', 'NaT'],
                          dtype='timedelta64[ns]')

        tdser = tm.box_expected(tdser, box_with_array)
        expected = tm.box_expected(expected, xbox)

        result = tdser * vector
        tm.assert_equal(result, expected)

        result = vector * tdser
        tm.assert_equal(result, expected)

    @pytest.mark.parametrize('dtype', ['int64', 'int32', 'int16',
                                       'uint64', 'uint32', 'uint16', 'uint8',
                                       'float64', 'float32', 'float16'])
    @pytest.mark.parametrize('vector', [np.array([20, 30, 40]),
                                        pd.Index([20, 30, 40]),
                                        Series([20, 30, 40])],
                             ids=lambda x: type(x).__name__)
    def test_td64arr_div_numeric_array(self, box_with_array, vector, dtype):
        # GH#4521
        # divide/multiply by integers
        xbox = get_upcast_box(box_with_array, vector)
        tdser = pd.Series(['59 Days', '59 Days', 'NaT'], dtype='m8[ns]')
        vector = vector.astype(dtype)
        expected = Series(['2.95D', '1D 23H 12m', 'NaT'],
                          dtype='timedelta64[ns]')

        tdser = tm.box_expected(tdser, box_with_array)
        expected = tm.box_expected(expected, xbox)

        result = tdser / vector
        tm.assert_equal(result, expected)

        pattern = ('true_divide cannot use operands|'
                   'cannot perform __div__|'
                   'cannot perform __truediv__|'
                   'unsupported operand|'
                   'Cannot divide')
        with pytest.raises(TypeError, match=pattern):
            vector / tdser

        if not isinstance(vector, pd.Index):
            # Index.__rdiv__ won't try to operate elementwise, just raises
            result = tdser / vector.astype(object)
            if box_with_array is pd.DataFrame:
                expected = [tdser.iloc[0, n] / vector[n]
                            for n in range(len(vector))]
            else:
                expected = [tdser[n] / vector[n] for n in range(len(tdser))]
            expected = tm.box_expected(expected, xbox)
            tm.assert_equal(result, expected)

        with pytest.raises(TypeError, match=pattern):
            vector.astype(object) / tdser

    @pytest.mark.parametrize('names', [(None, None, None),
                                       ('Egon', 'Venkman', None),
                                       ('NCC1701D', 'NCC1701D', 'NCC1701D')])
    def test_td64arr_mul_int_series(self, box_df_fail, names):
        # GH#19042 test for correct name attachment
        box = box_df_fail  # broadcasts along wrong axis, but doesn't raise
        tdi = TimedeltaIndex(['0days', '1day', '2days', '3days', '4days'],
                             name=names[0])
        # TODO: Should we be parametrizing over types for `ser` too?
        ser = Series([0, 1, 2, 3, 4], dtype=np.int64, name=names[1])

        expected = Series(['0days', '1day', '4days', '9days', '16days'],
                          dtype='timedelta64[ns]',
                          name=names[2])

        tdi = tm.box_expected(tdi, box)
        box = Series if (box is pd.Index and type(ser) is Series) else box
        expected = tm.box_expected(expected, box)

        result = ser * tdi
        tm.assert_equal(result, expected)

        # The direct operation tdi * ser still needs to be fixed.
        result = ser.__rmul__(tdi)
        tm.assert_equal(result, expected)

    # TODO: Should we be parametrizing over types for `ser` too?
    @pytest.mark.parametrize('names', [(None, None, None),
                                       ('Egon', 'Venkman', None),
                                       ('NCC1701D', 'NCC1701D', 'NCC1701D')])
    def test_float_series_rdiv_td64arr(self, box_with_array, names):
        # GH#19042 test for correct name attachment
        # TODO: the direct operation TimedeltaIndex / Series still
        # needs to be fixed.
        box = box_with_array
        tdi = TimedeltaIndex(['0days', '1day', '2days', '3days', '4days'],
                             name=names[0])
        ser = Series([1.5, 3, 4.5, 6, 7.5], dtype=np.float64, name=names[1])

        xname = names[2] if box is not tm.to_array else names[1]
        expected = Series([tdi[n] / ser[n] for n in range(len(ser))],
                          dtype='timedelta64[ns]',
                          name=xname)

        xbox = box
        if box in [pd.Index, tm.to_array] and type(ser) is Series:
            xbox = Series

        tdi = tm.box_expected(tdi, box)
        expected = tm.box_expected(expected, xbox)

        result = ser.__rdiv__(tdi)
        if box is pd.DataFrame:
            # TODO: Should we skip this case sooner or test something else?
            assert result is NotImplemented
        else:
            tm.assert_equal(result, expected)


class TestTimedeltaArraylikeInvalidArithmeticOps(object):

    def test_td64arr_pow_invalid(self, scalar_td, box_with_array):
        td1 = Series([timedelta(minutes=5, seconds=3)] * 3)
        td1.iloc[2] = np.nan

        td1 = tm.box_expected(td1, box_with_array)

        # check that we are getting a TypeError
        # with 'operate' (from core/ops.py) for the ops that are not
        # defined
        pattern = 'operate|unsupported|cannot|not supported'
        with pytest.raises(TypeError, match=pattern):
            scalar_td ** td1

        with pytest.raises(TypeError, match=pattern):
            td1 ** scalar_td